# Qwen Flash-Next: TP2 concurrency and cold-prefix cohorts

The k76 v2.0.0 concurrency series uses the earlier 16c9bd54788d12390838a65ce4a4ecda97fa5f1d GDN-MSE checkpoint, while v2.2.0 cold-prefix measurements use 03f4a0570496bbe741189f4a2a7d611679e85406. Both employ two directly connected Sparks, vLLM plus b12x and mixed NVFP4/MXFP8 paths. Release image digests and abbreviated engine commits do not establish an exact executed image digest.

- Type: Performance report
- Evidence: source_reported; site reproduced: false
- Model family: Qwen
- Model: ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE
- Original author: ursuciprian
- Original project: https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2
- Pinned document: https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/README.md
- Source revision: b384f4a0e3ff0d563980befe52eb3e6964a5b6b1
- Source commit time (not experiment date): 2026-10-17T10:48:19Z
- Checked: 2026-10-18
- Experiment date: 2026-10-10
- Runtime topics: vLLM
- Hardware topics: spark, nvidia
- Reported setup: {"system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled", "node_count": 2, "parallelism": "TP2", "kv_dtype": "fp8", "context_limit_tokens": 262144, "experiment_date": "2026-10-10"}
- Performance figures: attributed selections with per-point conditions; no cross-source ranking

## Conditions

Unknown is null. Recipe options are not measured benchmark settings.

```json
{
  "conditions": {
    "model_revision": null,
    "weight_hash": null,
    "runtime_execution_revision": null,
    "input_tokens": null,
    "output_tokens": null,
    "concurrency": null,
    "cache_state": null,
    "quantization": null,
    "speculative_decoding": null,
    "temperature": null,
    "top_p": null
  },
  "test_profiles": [
    {
      "label": "TP2 v2.0.0 C1 2048 per_request_decode_rate",
      "evidence_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv#L1142",
      "conditions": {
        "model": "Qwen3.8-Flash-Next",
        "model_artifact": "ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE",
        "engine": "vLLM + b12x",
        "engine_version": null,
        "runtime_execution_revision": null,
        "executed_image_digest": null,
        "system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled",
        "node_count": 2,
        "parallelism": "TP2",
        "weight_precision": "NVFP4 routed experts and decode GDN; MXFP8 dense/attention and prefill GDN copies",
        "extra_prefill_revision": "7c4f1bc1a2d6847e0cbc01ac6b823f00251de8dd",
        "kv_dtype": "fp8",
        "draft_tokens": 4,
        "context_limit_tokens": 262144,
        "configured_max_requests": 16,
        "actual_output_tokens_per_request": null,
        "GPU_clock_cap_MHz": null,
        "model_revision": "16c9bd54788d12390838a65ce4a4ecda97fa5f1d",
        "recipe_release": "v2.0.0",
        "experiment_date": "2026-10-10",
        "harness": "llama-benchy 0.4.1.dev4+g0d4de4271",
        "source_run": "2x:results/k76-capability-matrix-20261010-1059/2x/d0.json",
        "source_metric": "tg_req",
        "source_campaign": "k76-capability-matrix",
        "concurrency": 1,
        "input_tokens": 2048,
        "output_tokens": 512,
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 20,
        "thinking": true,
        "cache_state": "prefix caching enabled; fresh 2048-token task, not a cold-cache campaign",
        "repetitions": "mean ± sd of 3 runs, one boot",
        "aggregate_denominator": "all output tokens / wall time including prompt processing",
        "per_request_denominator": "one output stream after first token",
        "release_image_digest": "sha256:8abb60eeeb79ca13680341c17d87975801da1218c06df2505856b23da7ed1821",
        "release_digest_is_proven_executed_digest": false
      }
    },
    {
      "label": "TP2 v2.0.0 C4 2048 per_request_decode_rate",
      "evidence_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv#L1146",
      "conditions": {
        "model": "Qwen3.8-Flash-Next",
        "model_artifact": "ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE",
        "engine": "vLLM + b12x",
        "engine_version": null,
        "runtime_execution_revision": null,
        "executed_image_digest": null,
        "system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled",
        "node_count": 2,
        "parallelism": "TP2",
        "weight_precision": "NVFP4 routed experts and decode GDN; MXFP8 dense/attention and prefill GDN copies",
        "extra_prefill_revision": "7c4f1bc1a2d6847e0cbc01ac6b823f00251de8dd",
        "kv_dtype": "fp8",
        "draft_tokens": 4,
        "context_limit_tokens": 262144,
        "configured_max_requests": 16,
        "actual_output_tokens_per_request": null,
        "GPU_clock_cap_MHz": null,
        "model_revision": "16c9bd54788d12390838a65ce4a4ecda97fa5f1d",
        "recipe_release": "v2.0.0",
        "experiment_date": "2026-10-10",
        "harness": "llama-benchy 0.4.1.dev4+g0d4de4271",
        "source_run": "2x:results/k76-capability-matrix-20261010-1059/2x/d0.json",
        "source_metric": "tg_req",
        "source_campaign": "k76-capability-matrix",
        "concurrency": 4,
        "input_tokens": 2048,
        "output_tokens": 512,
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 20,
        "thinking": true,
        "cache_state": "prefix caching enabled; fresh 2048-token task, not a cold-cache campaign",
        "repetitions": "mean ± sd of 3 runs, one boot",
        "aggregate_denominator": "all output tokens / wall time including prompt processing",
        "per_request_denominator": "one output stream after first token",
        "release_image_digest": "sha256:8abb60eeeb79ca13680341c17d87975801da1218c06df2505856b23da7ed1821",
        "release_digest_is_proven_executed_digest": false
      }
    },
    {
      "label": "TP2 v2.0.0 C8 2048 per_request_decode_rate",
      "evidence_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv#L1148",
      "conditions": {
        "model": "Qwen3.8-Flash-Next",
        "model_artifact": "ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE",
        "engine": "vLLM + b12x",
        "engine_version": null,
        "runtime_execution_revision": null,
        "executed_image_digest": null,
        "system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled",
        "node_count": 2,
        "parallelism": "TP2",
        "weight_precision": "NVFP4 routed experts and decode GDN; MXFP8 dense/attention and prefill GDN copies",
        "extra_prefill_revision": "7c4f1bc1a2d6847e0cbc01ac6b823f00251de8dd",
        "kv_dtype": "fp8",
        "draft_tokens": 4,
        "context_limit_tokens": 262144,
        "configured_max_requests": 16,
        "actual_output_tokens_per_request": null,
        "GPU_clock_cap_MHz": null,
        "model_revision": "16c9bd54788d12390838a65ce4a4ecda97fa5f1d",
        "recipe_release": "v2.0.0",
        "experiment_date": "2026-10-10",
        "harness": "llama-benchy 0.4.1.dev4+g0d4de4271",
        "source_run": "2x:results/k76-capability-matrix-20261010-1059/2x/d0.json",
        "source_metric": "tg_req",
        "source_campaign": "k76-capability-matrix",
        "concurrency": 8,
        "input_tokens": 2048,
        "output_tokens": 512,
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 20,
        "thinking": true,
        "cache_state": "prefix caching enabled; fresh 2048-token task, not a cold-cache campaign",
        "repetitions": "mean ± sd of 3 runs, one boot",
        "aggregate_denominator": "all output tokens / wall time including prompt processing",
        "per_request_denominator": "one output stream after first token",
        "release_image_digest": "sha256:8abb60eeeb79ca13680341c17d87975801da1218c06df2505856b23da7ed1821",
        "release_digest_is_proven_executed_digest": false
      }
    },
    {
      "label": "TP2 v2.0.0 C16 2048 per_request_decode_rate",
      "evidence_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv#L1150",
      "conditions": {
        "model": "Qwen3.8-Flash-Next",
        "model_artifact": "ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE",
        "engine": "vLLM + b12x",
        "engine_version": null,
        "runtime_execution_revision": null,
        "executed_image_digest": null,
        "system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled",
        "node_count": 2,
        "parallelism": "TP2",
        "weight_precision": "NVFP4 routed experts and decode GDN; MXFP8 dense/attention and prefill GDN copies",
        "extra_prefill_revision": "7c4f1bc1a2d6847e0cbc01ac6b823f00251de8dd",
        "kv_dtype": "fp8",
        "draft_tokens": 4,
        "context_limit_tokens": 262144,
        "configured_max_requests": 16,
        "actual_output_tokens_per_request": null,
        "GPU_clock_cap_MHz": null,
        "model_revision": "16c9bd54788d12390838a65ce4a4ecda97fa5f1d",
        "recipe_release": "v2.0.0",
        "experiment_date": "2026-10-10",
        "harness": "llama-benchy 0.4.1.dev4+g0d4de4271",
        "source_run": "2x:results/k76-capability-matrix-20261010-1059/2x/d0.json",
        "source_metric": "tg_req",
        "source_campaign": "k76-capability-matrix",
        "concurrency": 16,
        "input_tokens": 2048,
        "output_tokens": 512,
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 20,
        "thinking": true,
        "cache_state": "prefix caching enabled; fresh 2048-token task, not a cold-cache campaign",
        "repetitions": "mean ± sd of 3 runs, one boot",
        "aggregate_denominator": "all output tokens / wall time including prompt processing",
        "per_request_denominator": "one output stream after first token",
        "release_image_digest": "sha256:8abb60eeeb79ca13680341c17d87975801da1218c06df2505856b23da7ed1821",
        "release_digest_is_proven_executed_digest": false
      }
    },
    {
      "label": "TP2 v2.0.0 C1 2048 aggregate_output_rate",
      "evidence_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv#L1152",
      "conditions": {
        "model": "Qwen3.8-Flash-Next",
        "model_artifact": "ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE",
        "engine": "vLLM + b12x",
        "engine_version": null,
        "runtime_execution_revision": null,
        "executed_image_digest": null,
        "system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled",
        "node_count": 2,
        "parallelism": "TP2",
        "weight_precision": "NVFP4 routed experts and decode GDN; MXFP8 dense/attention and prefill GDN copies",
        "extra_prefill_revision": "7c4f1bc1a2d6847e0cbc01ac6b823f00251de8dd",
        "kv_dtype": "fp8",
        "draft_tokens": 4,
        "context_limit_tokens": 262144,
        "configured_max_requests": 16,
        "actual_output_tokens_per_request": null,
        "GPU_clock_cap_MHz": null,
        "model_revision": "16c9bd54788d12390838a65ce4a4ecda97fa5f1d",
        "recipe_release": "v2.0.0",
        "experiment_date": "2026-10-10",
        "harness": "llama-benchy 0.4.1.dev4+g0d4de4271",
        "source_run": "2x:results/k76-capability-matrix-20261010-1059/2x/d0.json",
        "source_metric": "tg_total",
        "source_campaign": "k76-capability-matrix",
        "concurrency": 1,
        "input_tokens": 2048,
        "output_tokens": 512,
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 20,
        "thinking": true,
        "cache_state": "prefix caching enabled; fresh 2048-token task, not a cold-cache campaign",
        "repetitions": "mean ± sd of 3 runs, one boot",
        "aggregate_denominator": "all output tokens / wall time including prompt processing",
        "per_request_denominator": "one output stream after first token",
        "release_image_digest": "sha256:8abb60eeeb79ca13680341c17d87975801da1218c06df2505856b23da7ed1821",
        "release_digest_is_proven_executed_digest": false
      }
    },
    {
      "label": "TP2 v2.0.0 C4 2048 aggregate_output_rate",
      "evidence_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv#L1156",
      "conditions": {
        "model": "Qwen3.8-Flash-Next",
        "model_artifact": "ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE",
        "engine": "vLLM + b12x",
        "engine_version": null,
        "runtime_execution_revision": null,
        "executed_image_digest": null,
        "system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled",
        "node_count": 2,
        "parallelism": "TP2",
        "weight_precision": "NVFP4 routed experts and decode GDN; MXFP8 dense/attention and prefill GDN copies",
        "extra_prefill_revision": "7c4f1bc1a2d6847e0cbc01ac6b823f00251de8dd",
        "kv_dtype": "fp8",
        "draft_tokens": 4,
        "context_limit_tokens": 262144,
        "configured_max_requests": 16,
        "actual_output_tokens_per_request": null,
        "GPU_clock_cap_MHz": null,
        "model_revision": "16c9bd54788d12390838a65ce4a4ecda97fa5f1d",
        "recipe_release": "v2.0.0",
        "experiment_date": "2026-10-10",
        "harness": "llama-benchy 0.4.1.dev4+g0d4de4271",
        "source_run": "2x:results/k76-capability-matrix-20261010-1059/2x/d0.json",
        "source_metric": "tg_total",
        "source_campaign": "k76-capability-matrix",
        "concurrency": 4,
        "input_tokens": 2048,
        "output_tokens": 512,
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 20,
        "thinking": true,
        "cache_state": "prefix caching enabled; fresh 2048-token task, not a cold-cache campaign",
        "repetitions": "mean ± sd of 3 runs, one boot",
        "aggregate_denominator": "all output tokens / wall time including prompt processing",
        "per_request_denominator": "one output stream after first token",
        "release_image_digest": "sha256:8abb60eeeb79ca13680341c17d87975801da1218c06df2505856b23da7ed1821",
        "release_digest_is_proven_executed_digest": false
      }
    },
    {
      "label": "TP2 v2.0.0 C8 2048 aggregate_output_rate",
      "evidence_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv#L1158",
      "conditions": {
        "model": "Qwen3.8-Flash-Next",
        "model_artifact": "ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE",
        "engine": "vLLM + b12x",
        "engine_version": null,
        "runtime_execution_revision": null,
        "executed_image_digest": null,
        "system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled",
        "node_count": 2,
        "parallelism": "TP2",
        "weight_precision": "NVFP4 routed experts and decode GDN; MXFP8 dense/attention and prefill GDN copies",
        "extra_prefill_revision": "7c4f1bc1a2d6847e0cbc01ac6b823f00251de8dd",
        "kv_dtype": "fp8",
        "draft_tokens": 4,
        "context_limit_tokens": 262144,
        "configured_max_requests": 16,
        "actual_output_tokens_per_request": null,
        "GPU_clock_cap_MHz": null,
        "model_revision": "16c9bd54788d12390838a65ce4a4ecda97fa5f1d",
        "recipe_release": "v2.0.0",
        "experiment_date": "2026-10-10",
        "harness": "llama-benchy 0.4.1.dev4+g0d4de4271",
        "source_run": "2x:results/k76-capability-matrix-20261010-1059/2x/d0.json",
        "source_metric": "tg_total",
        "source_campaign": "k76-capability-matrix",
        "concurrency": 8,
        "input_tokens": 2048,
        "output_tokens": 512,
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 20,
        "thinking": true,
        "cache_state": "prefix caching enabled; fresh 2048-token task, not a cold-cache campaign",
        "repetitions": "mean ± sd of 3 runs, one boot",
        "aggregate_denominator": "all output tokens / wall time including prompt processing",
        "per_request_denominator": "one output stream after first token",
        "release_image_digest": "sha256:8abb60eeeb79ca13680341c17d87975801da1218c06df2505856b23da7ed1821",
        "release_digest_is_proven_executed_digest": false
      }
    },
    {
      "label": "TP2 v2.0.0 C16 2048 aggregate_output_rate",
      "evidence_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv#L1160",
      "conditions": {
        "model": "Qwen3.8-Flash-Next",
        "model_artifact": "ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE",
        "engine": "vLLM + b12x",
        "engine_version": null,
        "runtime_execution_revision": null,
        "executed_image_digest": null,
        "system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled",
        "node_count": 2,
        "parallelism": "TP2",
        "weight_precision": "NVFP4 routed experts and decode GDN; MXFP8 dense/attention and prefill GDN copies",
        "extra_prefill_revision": "7c4f1bc1a2d6847e0cbc01ac6b823f00251de8dd",
        "kv_dtype": "fp8",
        "draft_tokens": 4,
        "context_limit_tokens": 262144,
        "configured_max_requests": 16,
        "actual_output_tokens_per_request": null,
        "GPU_clock_cap_MHz": null,
        "model_revision": "16c9bd54788d12390838a65ce4a4ecda97fa5f1d",
        "recipe_release": "v2.0.0",
        "experiment_date": "2026-10-10",
        "harness": "llama-benchy 0.4.1.dev4+g0d4de4271",
        "source_run": "2x:results/k76-capability-matrix-20261010-1059/2x/d0.json",
        "source_metric": "tg_total",
        "source_campaign": "k76-capability-matrix",
        "concurrency": 16,
        "input_tokens": 2048,
        "output_tokens": 512,
        "temperature": 1.0,
        "top_p": 0.95,
        "top_k": 20,
        "thinking": true,
        "cache_state": "prefix caching enabled; fresh 2048-token task, not a cold-cache campaign",
        "repetitions": "mean ± sd of 3 runs, one boot",
        "aggregate_denominator": "all output tokens / wall time including prompt processing",
        "per_request_denominator": "one output stream after first token",
        "release_image_digest": "sha256:8abb60eeeb79ca13680341c17d87975801da1218c06df2505856b23da7ed1821",
        "release_digest_is_proven_executed_digest": false
      }
    },
    {
      "label": "TP2 v2.2.0 C1 16384 ttft",
      "evidence_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv#L1262",
      "conditions": {
        "model": "Qwen3.8-Flash-Next",
        "model_artifact": "ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE",
        "engine": "vLLM + b12x",
        "engine_version": null,
        "runtime_execution_revision": null,
        "executed_image_digest": null,
        "system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled",
        "node_count": 2,
        "parallelism": "TP2",
        "weight_precision": "NVFP4 routed experts and decode GDN; MXFP8 dense/attention and prefill GDN copies",
        "extra_prefill_revision": "7c4f1bc1a2d6847e0cbc01ac6b823f00251de8dd",
        "kv_dtype": "fp8",
        "draft_tokens": 4,
        "context_limit_tokens": 262144,
        "configured_max_requests": 16,
        "actual_output_tokens_per_request": null,
        "GPU_clock_cap_MHz": null,
        "model_revision": "03f4a0570496bbe741189f4a2a7d611679e85406",
        "recipe_release": "v2.2.0",
        "experiment_date": "2026-10-15",
        "harness": "cold.py (k101)",
        "source_run": "1x:results/k101s-1x-gdn-prefill-views-ab-20261015-1600/tp2/2x/",
        "source_metric": "ttft_s",
        "source_campaign": "promotion A/B",
        "concurrency": 1,
        "input_tokens": 16384,
        "output_tokens": 1,
        "temperature": 0,
        "top_p": null,
        "top_k": null,
        "thinking": null,
        "cache_state": "cold prefix with no hit, not a cold OS page-cache claim",
        "repetitions": "mean of 4 boots x 3 cold prompts, sd between boots",
        "aggregate_denominator": null,
        "per_request_denominator": null,
        "release_image_digest": null,
        "release_digest_is_proven_executed_digest": false
      }
    },
    {
      "label": "TP2 v2.2.0 C1 16384 prefill_rate",
      "evidence_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv#L1263",
      "conditions": {
        "model": "Qwen3.8-Flash-Next",
        "model_artifact": "ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE",
        "engine": "vLLM + b12x",
        "engine_version": null,
        "runtime_execution_revision": null,
        "executed_image_digest": null,
        "system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled",
        "node_count": 2,
        "parallelism": "TP2",
        "weight_precision": "NVFP4 routed experts and decode GDN; MXFP8 dense/attention and prefill GDN copies",
        "extra_prefill_revision": "7c4f1bc1a2d6847e0cbc01ac6b823f00251de8dd",
        "kv_dtype": "fp8",
        "draft_tokens": 4,
        "context_limit_tokens": 262144,
        "configured_max_requests": 16,
        "actual_output_tokens_per_request": null,
        "GPU_clock_cap_MHz": null,
        "model_revision": "03f4a0570496bbe741189f4a2a7d611679e85406",
        "recipe_release": "v2.2.0",
        "experiment_date": "2026-10-15",
        "harness": "cold.py (k101)",
        "source_run": "1x:results/k101s-1x-gdn-prefill-views-ab-20261015-1600/tp2/2x/",
        "source_metric": "pp",
        "source_campaign": "promotion A/B",
        "concurrency": 1,
        "input_tokens": 16384,
        "output_tokens": 1,
        "temperature": 0,
        "top_p": null,
        "top_k": null,
        "thinking": null,
        "cache_state": "cold prefix with no hit, not a cold OS page-cache claim",
        "repetitions": "mean of 4 boots x 3 cold prompts, sd between boots",
        "aggregate_denominator": null,
        "per_request_denominator": null,
        "release_image_digest": null,
        "release_digest_is_proven_executed_digest": false
      }
    },
    {
      "label": "TP2 v2.2.0 C1 65536 ttft",
      "evidence_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv#L1264",
      "conditions": {
        "model": "Qwen3.8-Flash-Next",
        "model_artifact": "ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE",
        "engine": "vLLM + b12x",
        "engine_version": null,
        "runtime_execution_revision": null,
        "executed_image_digest": null,
        "system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled",
        "node_count": 2,
        "parallelism": "TP2",
        "weight_precision": "NVFP4 routed experts and decode GDN; MXFP8 dense/attention and prefill GDN copies",
        "extra_prefill_revision": "7c4f1bc1a2d6847e0cbc01ac6b823f00251de8dd",
        "kv_dtype": "fp8",
        "draft_tokens": 4,
        "context_limit_tokens": 262144,
        "configured_max_requests": 16,
        "actual_output_tokens_per_request": null,
        "GPU_clock_cap_MHz": null,
        "model_revision": "03f4a0570496bbe741189f4a2a7d611679e85406",
        "recipe_release": "v2.2.0",
        "experiment_date": "2026-10-15",
        "harness": "cold.py (k101)",
        "source_run": "1x:results/k101s-1x-gdn-prefill-views-ab-20261015-1600/tp2/2x/",
        "source_metric": "ttft_s",
        "source_campaign": "promotion A/B",
        "concurrency": 1,
        "input_tokens": 65536,
        "output_tokens": 1,
        "temperature": 0,
        "top_p": null,
        "top_k": null,
        "thinking": null,
        "cache_state": "cold prefix with no hit, not a cold OS page-cache claim",
        "repetitions": "mean of 4 boots x 3 cold prompts, sd between boots",
        "aggregate_denominator": null,
        "per_request_denominator": null,
        "release_image_digest": null,
        "release_digest_is_proven_executed_digest": false
      }
    },
    {
      "label": "TP2 v2.2.0 C1 65536 prefill_rate",
      "evidence_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv#L1265",
      "conditions": {
        "model": "Qwen3.8-Flash-Next",
        "model_artifact": "ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE",
        "engine": "vLLM + b12x",
        "engine_version": null,
        "runtime_execution_revision": null,
        "executed_image_digest": null,
        "system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled",
        "node_count": 2,
        "parallelism": "TP2",
        "weight_precision": "NVFP4 routed experts and decode GDN; MXFP8 dense/attention and prefill GDN copies",
        "extra_prefill_revision": "7c4f1bc1a2d6847e0cbc01ac6b823f00251de8dd",
        "kv_dtype": "fp8",
        "draft_tokens": 4,
        "context_limit_tokens": 262144,
        "configured_max_requests": 16,
        "actual_output_tokens_per_request": null,
        "GPU_clock_cap_MHz": null,
        "model_revision": "03f4a0570496bbe741189f4a2a7d611679e85406",
        "recipe_release": "v2.2.0",
        "experiment_date": "2026-10-15",
        "harness": "cold.py (k101)",
        "source_run": "1x:results/k101s-1x-gdn-prefill-views-ab-20261015-1600/tp2/2x/",
        "source_metric": "pp",
        "source_campaign": "promotion A/B",
        "concurrency": 1,
        "input_tokens": 65536,
        "output_tokens": 1,
        "temperature": 0,
        "top_p": null,
        "top_k": null,
        "thinking": null,
        "cache_state": "cold prefix with no hit, not a cold OS page-cache claim",
        "repetitions": "mean of 4 boots x 3 cold prompts, sd between boots",
        "aggregate_denominator": null,
        "per_request_denominator": null,
        "release_image_digest": null,
        "release_digest_is_proven_executed_digest": false
      }
    },
    {
      "label": "TP2 v2.2.0 C1 131072 ttft",
      "evidence_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv#L1266",
      "conditions": {
        "model": "Qwen3.8-Flash-Next",
        "model_artifact": "ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE",
        "engine": "vLLM + b12x",
        "engine_version": null,
        "runtime_execution_revision": null,
        "executed_image_digest": null,
        "system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled",
        "node_count": 2,
        "parallelism": "TP2",
        "weight_precision": "NVFP4 routed experts and decode GDN; MXFP8 dense/attention and prefill GDN copies",
        "extra_prefill_revision": "7c4f1bc1a2d6847e0cbc01ac6b823f00251de8dd",
        "kv_dtype": "fp8",
        "draft_tokens": 4,
        "context_limit_tokens": 262144,
        "configured_max_requests": 16,
        "actual_output_tokens_per_request": null,
        "GPU_clock_cap_MHz": null,
        "model_revision": "03f4a0570496bbe741189f4a2a7d611679e85406",
        "recipe_release": "v2.2.0",
        "experiment_date": "2026-10-15",
        "harness": "cold.py (k101)",
        "source_run": "1x:results/k101s-1x-gdn-prefill-views-ab-20261015-1600/tp2/2x/",
        "source_metric": "ttft_s",
        "source_campaign": "promotion A/B",
        "concurrency": 1,
        "input_tokens": 131072,
        "output_tokens": 1,
        "temperature": 0,
        "top_p": null,
        "top_k": null,
        "thinking": null,
        "cache_state": "cold prefix with no hit, not a cold OS page-cache claim",
        "repetitions": "mean of 4 boots x 3 cold prompts, sd between boots",
        "aggregate_denominator": null,
        "per_request_denominator": null,
        "release_image_digest": null,
        "release_digest_is_proven_executed_digest": false
      }
    },
    {
      "label": "TP2 v2.2.0 C1 131072 prefill_rate",
      "evidence_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv#L1267",
      "conditions": {
        "model": "Qwen3.8-Flash-Next",
        "model_artifact": "ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE",
        "engine": "vLLM + b12x",
        "engine_version": null,
        "runtime_execution_revision": null,
        "executed_image_digest": null,
        "system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled",
        "node_count": 2,
        "parallelism": "TP2",
        "weight_precision": "NVFP4 routed experts and decode GDN; MXFP8 dense/attention and prefill GDN copies",
        "extra_prefill_revision": "7c4f1bc1a2d6847e0cbc01ac6b823f00251de8dd",
        "kv_dtype": "fp8",
        "draft_tokens": 4,
        "context_limit_tokens": 262144,
        "configured_max_requests": 16,
        "actual_output_tokens_per_request": null,
        "GPU_clock_cap_MHz": null,
        "model_revision": "03f4a0570496bbe741189f4a2a7d611679e85406",
        "recipe_release": "v2.2.0",
        "experiment_date": "2026-10-15",
        "harness": "cold.py (k101)",
        "source_run": "1x:results/k101s-1x-gdn-prefill-views-ab-20261015-1600/tp2/2x/",
        "source_metric": "pp",
        "source_campaign": "promotion A/B",
        "concurrency": 1,
        "input_tokens": 131072,
        "output_tokens": 1,
        "temperature": 0,
        "top_p": null,
        "top_k": null,
        "thinking": null,
        "cache_state": "cold prefix with no hit, not a cold OS page-cache claim",
        "repetitions": "mean of 4 boots x 3 cold prompts, sd between boots",
        "aggregate_denominator": null,
        "per_request_denominator": null,
        "release_image_digest": null,
        "release_digest_is_proven_executed_digest": false
      }
    }
  ],
  "recipe_parameters": null,
  "deployment_options": []
}
```

## Reading notes

- One source document can contain different experiments and versions; not an independent experiment count.

## Limits

- No independent site model execution.
- Missing sampling, exact executed image digests and measurement boundaries remain null.
- Configured window, completed input, scheduler ceiling and measured full-window concurrency are separate quantities.

## License declaration

{"declared_spdx": "apache-2.0", "scope": "Source declaration; models, runtime and dependencies retain their exact terms."}


## Editorial answer

The k76 v2.0.0 concurrency series uses the earlier 16c9bd54788d12390838a65ce4a4ecda97fa5f1d GDN-MSE checkpoint, while v2.2.0 cold-prefix measurements use 03f4a0570496bbe741189f4a2a7d611679e85406. Both employ two directly connected Sparks, vLLM plus b12x and mixed NVFP4/MXFP8 paths. Release image digests and abbreviated engine commits do not establish an exact executed image digest.

By Run AI Here editorial · 2026-10-18

Original source-based interpretation; local Spark Qwen draft with Codex factual review, no site model benchmark.

## Why keep releases separate?

The k76 v2.0.0 concurrency series uses the earlier 16c9bd54788d12390838a65ce4a4ecda97fa5f1d GDN-MSE checkpoint, while v2.2.0 cold-prefix measurements use 03f4a0570496bbe741189f4a2a7d611679e85406. Both employ two directly connected Sparks, vLLM plus b12x and mixed NVFP4/MXFP8 paths. Release image digests and abbreviated engine commits do not establish an exact executed image digest.

Sources: https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/VERSIONS.md

## What does the k76 concurrency series measure?

The selected cells use llama-benchy 0.4.1.dev4+g0d4de4271, 2048 input, 512 requested output tokens, temperature 1.0, top-p 0.95, top-k 20 and thinking on. Prefix caching is enabled. Means and SD cover three runs in one boot, not independent boots. Aggregate output includes prompt-processing wall time; individual decode begins after the first token.

Sources: https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv, https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/BENCHMARKS.md

## What does the later cold-prefix arm measure?

At C1 and temperature zero, v2.2.0 requests one output token after 16384, 65536 or 131072 input tokens. Four boots each contain three cold prompts. The SD describes variation between boot means. These observations measure first-token latency and prefill, not sustained generation, broad long-document accuracy or multi-user full-window serving.

Sources: https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv, https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/BENCHMARKS.md

## Which conclusions are outside these data?

The two arms cannot be joined into a single unchanged deployment curve. Current request slots and KV capacity are configuration limits, and cold prefix does not mean cold filesystem pages or cold compilation state. Complete executed digests and some runtime identities remain null. The publisher is already represented by a separate one-Spark document in the catalog; this new document is not a new independent researcher.

Sources: https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/README.md, https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/VERSIONS.md

## Selected source-reported figures

Each point retains its own configuration, measurement unit and source. No site reproduction or cross-source ranking.

```json
[
  {
    "label": "TP2 v2.0.0 C1 2048 per_request_decode_rate",
    "metric": "per_request_decode_rate",
    "value": 88.697,
    "unit": "tokens/s",
    "statistic": "mean ± sd of 3 runs, one boot",
    "uncertainty": {
      "standard_deviation": 2.562,
      "kind": "SD across runs in one boot"
    },
    "conditions": {
      "model": "Qwen3.8-Flash-Next",
      "model_artifact": "ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE",
      "engine": "vLLM + b12x",
      "engine_version": null,
      "runtime_execution_revision": null,
      "executed_image_digest": null,
      "system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled",
      "node_count": 2,
      "parallelism": "TP2",
      "weight_precision": "NVFP4 routed experts and decode GDN; MXFP8 dense/attention and prefill GDN copies",
      "extra_prefill_revision": "7c4f1bc1a2d6847e0cbc01ac6b823f00251de8dd",
      "kv_dtype": "fp8",
      "draft_tokens": 4,
      "context_limit_tokens": 262144,
      "configured_max_requests": 16,
      "actual_output_tokens_per_request": null,
      "GPU_clock_cap_MHz": null,
      "model_revision": "16c9bd54788d12390838a65ce4a4ecda97fa5f1d",
      "recipe_release": "v2.0.0",
      "experiment_date": "2026-10-10",
      "harness": "llama-benchy 0.4.1.dev4+g0d4de4271",
      "source_run": "2x:results/k76-capability-matrix-20261010-1059/2x/d0.json",
      "source_metric": "tg_req",
      "source_campaign": "k76-capability-matrix",
      "concurrency": 1,
      "input_tokens": 2048,
      "output_tokens": 512,
      "temperature": 1.0,
      "top_p": 0.95,
      "top_k": 20,
      "thinking": true,
      "cache_state": "prefix caching enabled; fresh 2048-token task, not a cold-cache campaign",
      "repetitions": "mean ± sd of 3 runs, one boot",
      "aggregate_denominator": "all output tokens / wall time including prompt processing",
      "per_request_denominator": "one output stream after first token",
      "release_image_digest": "sha256:8abb60eeeb79ca13680341c17d87975801da1218c06df2505856b23da7ed1821",
      "release_digest_is_proven_executed_digest": false
    },
    "evidence_state": "source_reported",
    "site_reproduced": false,
    "source_document_id": "qwen-tp2-csv",
    "source_document_sha256": "a82bffe4f0e33f4243db9549c02fb8e9379e5faa62825a35a7b087cd30b4996e",
    "source_line": 1142,
    "evidence_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv#L1142"
  },
  {
    "label": "TP2 v2.0.0 C4 2048 per_request_decode_rate",
    "metric": "per_request_decode_rate",
    "value": 48.631,
    "unit": "tokens/s",
    "statistic": "mean ± sd of 3 runs, one boot",
    "uncertainty": {
      "standard_deviation": 1.533,
      "kind": "SD across runs in one boot"
    },
    "conditions": {
      "model": "Qwen3.8-Flash-Next",
      "model_artifact": "ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE",
      "engine": "vLLM + b12x",
      "engine_version": null,
      "runtime_execution_revision": null,
      "executed_image_digest": null,
      "system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled",
      "node_count": 2,
      "parallelism": "TP2",
      "weight_precision": "NVFP4 routed experts and decode GDN; MXFP8 dense/attention and prefill GDN copies",
      "extra_prefill_revision": "7c4f1bc1a2d6847e0cbc01ac6b823f00251de8dd",
      "kv_dtype": "fp8",
      "draft_tokens": 4,
      "context_limit_tokens": 262144,
      "configured_max_requests": 16,
      "actual_output_tokens_per_request": null,
      "GPU_clock_cap_MHz": null,
      "model_revision": "16c9bd54788d12390838a65ce4a4ecda97fa5f1d",
      "recipe_release": "v2.0.0",
      "experiment_date": "2026-10-10",
      "harness": "llama-benchy 0.4.1.dev4+g0d4de4271",
      "source_run": "2x:results/k76-capability-matrix-20261010-1059/2x/d0.json",
      "source_metric": "tg_req",
      "source_campaign": "k76-capability-matrix",
      "concurrency": 4,
      "input_tokens": 2048,
      "output_tokens": 512,
      "temperature": 1.0,
      "top_p": 0.95,
      "top_k": 20,
      "thinking": true,
      "cache_state": "prefix caching enabled; fresh 2048-token task, not a cold-cache campaign",
      "repetitions": "mean ± sd of 3 runs, one boot",
      "aggregate_denominator": "all output tokens / wall time including prompt processing",
      "per_request_denominator": "one output stream after first token",
      "release_image_digest": "sha256:8abb60eeeb79ca13680341c17d87975801da1218c06df2505856b23da7ed1821",
      "release_digest_is_proven_executed_digest": false
    },
    "evidence_state": "source_reported",
    "site_reproduced": false,
    "source_document_id": "qwen-tp2-csv",
    "source_document_sha256": "a82bffe4f0e33f4243db9549c02fb8e9379e5faa62825a35a7b087cd30b4996e",
    "source_line": 1146,
    "evidence_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv#L1146"
  },
  {
    "label": "TP2 v2.0.0 C8 2048 per_request_decode_rate",
    "metric": "per_request_decode_rate",
    "value": 33.236,
    "unit": "tokens/s",
    "statistic": "mean ± sd of 3 runs, one boot",
    "uncertainty": {
      "standard_deviation": 0.977,
      "kind": "SD across runs in one boot"
    },
    "conditions": {
      "model": "Qwen3.8-Flash-Next",
      "model_artifact": "ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE",
      "engine": "vLLM + b12x",
      "engine_version": null,
      "runtime_execution_revision": null,
      "executed_image_digest": null,
      "system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled",
      "node_count": 2,
      "parallelism": "TP2",
      "weight_precision": "NVFP4 routed experts and decode GDN; MXFP8 dense/attention and prefill GDN copies",
      "extra_prefill_revision": "7c4f1bc1a2d6847e0cbc01ac6b823f00251de8dd",
      "kv_dtype": "fp8",
      "draft_tokens": 4,
      "context_limit_tokens": 262144,
      "configured_max_requests": 16,
      "actual_output_tokens_per_request": null,
      "GPU_clock_cap_MHz": null,
      "model_revision": "16c9bd54788d12390838a65ce4a4ecda97fa5f1d",
      "recipe_release": "v2.0.0",
      "experiment_date": "2026-10-10",
      "harness": "llama-benchy 0.4.1.dev4+g0d4de4271",
      "source_run": "2x:results/k76-capability-matrix-20261010-1059/2x/d0.json",
      "source_metric": "tg_req",
      "source_campaign": "k76-capability-matrix",
      "concurrency": 8,
      "input_tokens": 2048,
      "output_tokens": 512,
      "temperature": 1.0,
      "top_p": 0.95,
      "top_k": 20,
      "thinking": true,
      "cache_state": "prefix caching enabled; fresh 2048-token task, not a cold-cache campaign",
      "repetitions": "mean ± sd of 3 runs, one boot",
      "aggregate_denominator": "all output tokens / wall time including prompt processing",
      "per_request_denominator": "one output stream after first token",
      "release_image_digest": "sha256:8abb60eeeb79ca13680341c17d87975801da1218c06df2505856b23da7ed1821",
      "release_digest_is_proven_executed_digest": false
    },
    "evidence_state": "source_reported",
    "site_reproduced": false,
    "source_document_id": "qwen-tp2-csv",
    "source_document_sha256": "a82bffe4f0e33f4243db9549c02fb8e9379e5faa62825a35a7b087cd30b4996e",
    "source_line": 1148,
    "evidence_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv#L1148"
  },
  {
    "label": "TP2 v2.0.0 C16 2048 per_request_decode_rate",
    "metric": "per_request_decode_rate",
    "value": 22.142,
    "unit": "tokens/s",
    "statistic": "mean ± sd of 3 runs, one boot",
    "uncertainty": {
      "standard_deviation": 0.362,
      "kind": "SD across runs in one boot"
    },
    "conditions": {
      "model": "Qwen3.8-Flash-Next",
      "model_artifact": "ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE",
      "engine": "vLLM + b12x",
      "engine_version": null,
      "runtime_execution_revision": null,
      "executed_image_digest": null,
      "system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled",
      "node_count": 2,
      "parallelism": "TP2",
      "weight_precision": "NVFP4 routed experts and decode GDN; MXFP8 dense/attention and prefill GDN copies",
      "extra_prefill_revision": "7c4f1bc1a2d6847e0cbc01ac6b823f00251de8dd",
      "kv_dtype": "fp8",
      "draft_tokens": 4,
      "context_limit_tokens": 262144,
      "configured_max_requests": 16,
      "actual_output_tokens_per_request": null,
      "GPU_clock_cap_MHz": null,
      "model_revision": "16c9bd54788d12390838a65ce4a4ecda97fa5f1d",
      "recipe_release": "v2.0.0",
      "experiment_date": "2026-10-10",
      "harness": "llama-benchy 0.4.1.dev4+g0d4de4271",
      "source_run": "2x:results/k76-capability-matrix-20261010-1059/2x/d0.json",
      "source_metric": "tg_req",
      "source_campaign": "k76-capability-matrix",
      "concurrency": 16,
      "input_tokens": 2048,
      "output_tokens": 512,
      "temperature": 1.0,
      "top_p": 0.95,
      "top_k": 20,
      "thinking": true,
      "cache_state": "prefix caching enabled; fresh 2048-token task, not a cold-cache campaign",
      "repetitions": "mean ± sd of 3 runs, one boot",
      "aggregate_denominator": "all output tokens / wall time including prompt processing",
      "per_request_denominator": "one output stream after first token",
      "release_image_digest": "sha256:8abb60eeeb79ca13680341c17d87975801da1218c06df2505856b23da7ed1821",
      "release_digest_is_proven_executed_digest": false
    },
    "evidence_state": "source_reported",
    "site_reproduced": false,
    "source_document_id": "qwen-tp2-csv",
    "source_document_sha256": "a82bffe4f0e33f4243db9549c02fb8e9379e5faa62825a35a7b087cd30b4996e",
    "source_line": 1150,
    "evidence_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv#L1150"
  },
  {
    "label": "TP2 v2.0.0 C1 2048 aggregate_output_rate",
    "metric": "aggregate_output_rate",
    "value": 88.697,
    "unit": "tokens/s",
    "statistic": "mean ± sd of 3 runs, one boot",
    "uncertainty": {
      "standard_deviation": 2.562,
      "kind": "SD across runs in one boot"
    },
    "conditions": {
      "model": "Qwen3.8-Flash-Next",
      "model_artifact": "ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE",
      "engine": "vLLM + b12x",
      "engine_version": null,
      "runtime_execution_revision": null,
      "executed_image_digest": null,
      "system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled",
      "node_count": 2,
      "parallelism": "TP2",
      "weight_precision": "NVFP4 routed experts and decode GDN; MXFP8 dense/attention and prefill GDN copies",
      "extra_prefill_revision": "7c4f1bc1a2d6847e0cbc01ac6b823f00251de8dd",
      "kv_dtype": "fp8",
      "draft_tokens": 4,
      "context_limit_tokens": 262144,
      "configured_max_requests": 16,
      "actual_output_tokens_per_request": null,
      "GPU_clock_cap_MHz": null,
      "model_revision": "16c9bd54788d12390838a65ce4a4ecda97fa5f1d",
      "recipe_release": "v2.0.0",
      "experiment_date": "2026-10-10",
      "harness": "llama-benchy 0.4.1.dev4+g0d4de4271",
      "source_run": "2x:results/k76-capability-matrix-20261010-1059/2x/d0.json",
      "source_metric": "tg_total",
      "source_campaign": "k76-capability-matrix",
      "concurrency": 1,
      "input_tokens": 2048,
      "output_tokens": 512,
      "temperature": 1.0,
      "top_p": 0.95,
      "top_k": 20,
      "thinking": true,
      "cache_state": "prefix caching enabled; fresh 2048-token task, not a cold-cache campaign",
      "repetitions": "mean ± sd of 3 runs, one boot",
      "aggregate_denominator": "all output tokens / wall time including prompt processing",
      "per_request_denominator": "one output stream after first token",
      "release_image_digest": "sha256:8abb60eeeb79ca13680341c17d87975801da1218c06df2505856b23da7ed1821",
      "release_digest_is_proven_executed_digest": false
    },
    "evidence_state": "source_reported",
    "site_reproduced": false,
    "source_document_id": "qwen-tp2-csv",
    "source_document_sha256": "a82bffe4f0e33f4243db9549c02fb8e9379e5faa62825a35a7b087cd30b4996e",
    "source_line": 1152,
    "evidence_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv#L1152"
  },
  {
    "label": "TP2 v2.0.0 C4 2048 aggregate_output_rate",
    "metric": "aggregate_output_rate",
    "value": 156.645,
    "unit": "tokens/s",
    "statistic": "mean ± sd of 3 runs, one boot",
    "uncertainty": {
      "standard_deviation": 4.558,
      "kind": "SD across runs in one boot"
    },
    "conditions": {
      "model": "Qwen3.8-Flash-Next",
      "model_artifact": "ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE",
      "engine": "vLLM + b12x",
      "engine_version": null,
      "runtime_execution_revision": null,
      "executed_image_digest": null,
      "system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled",
      "node_count": 2,
      "parallelism": "TP2",
      "weight_precision": "NVFP4 routed experts and decode GDN; MXFP8 dense/attention and prefill GDN copies",
      "extra_prefill_revision": "7c4f1bc1a2d6847e0cbc01ac6b823f00251de8dd",
      "kv_dtype": "fp8",
      "draft_tokens": 4,
      "context_limit_tokens": 262144,
      "configured_max_requests": 16,
      "actual_output_tokens_per_request": null,
      "GPU_clock_cap_MHz": null,
      "model_revision": "16c9bd54788d12390838a65ce4a4ecda97fa5f1d",
      "recipe_release": "v2.0.0",
      "experiment_date": "2026-10-10",
      "harness": "llama-benchy 0.4.1.dev4+g0d4de4271",
      "source_run": "2x:results/k76-capability-matrix-20261010-1059/2x/d0.json",
      "source_metric": "tg_total",
      "source_campaign": "k76-capability-matrix",
      "concurrency": 4,
      "input_tokens": 2048,
      "output_tokens": 512,
      "temperature": 1.0,
      "top_p": 0.95,
      "top_k": 20,
      "thinking": true,
      "cache_state": "prefix caching enabled; fresh 2048-token task, not a cold-cache campaign",
      "repetitions": "mean ± sd of 3 runs, one boot",
      "aggregate_denominator": "all output tokens / wall time including prompt processing",
      "per_request_denominator": "one output stream after first token",
      "release_image_digest": "sha256:8abb60eeeb79ca13680341c17d87975801da1218c06df2505856b23da7ed1821",
      "release_digest_is_proven_executed_digest": false
    },
    "evidence_state": "source_reported",
    "site_reproduced": false,
    "source_document_id": "qwen-tp2-csv",
    "source_document_sha256": "a82bffe4f0e33f4243db9549c02fb8e9379e5faa62825a35a7b087cd30b4996e",
    "source_line": 1156,
    "evidence_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv#L1156"
  },
  {
    "label": "TP2 v2.0.0 C8 2048 aggregate_output_rate",
    "metric": "aggregate_output_rate",
    "value": 172.213,
    "unit": "tokens/s",
    "statistic": "mean ± sd of 3 runs, one boot",
    "uncertainty": {
      "standard_deviation": 16.348,
      "kind": "SD across runs in one boot"
    },
    "conditions": {
      "model": "Qwen3.8-Flash-Next",
      "model_artifact": "ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE",
      "engine": "vLLM + b12x",
      "engine_version": null,
      "runtime_execution_revision": null,
      "executed_image_digest": null,
      "system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled",
      "node_count": 2,
      "parallelism": "TP2",
      "weight_precision": "NVFP4 routed experts and decode GDN; MXFP8 dense/attention and prefill GDN copies",
      "extra_prefill_revision": "7c4f1bc1a2d6847e0cbc01ac6b823f00251de8dd",
      "kv_dtype": "fp8",
      "draft_tokens": 4,
      "context_limit_tokens": 262144,
      "configured_max_requests": 16,
      "actual_output_tokens_per_request": null,
      "GPU_clock_cap_MHz": null,
      "model_revision": "16c9bd54788d12390838a65ce4a4ecda97fa5f1d",
      "recipe_release": "v2.0.0",
      "experiment_date": "2026-10-10",
      "harness": "llama-benchy 0.4.1.dev4+g0d4de4271",
      "source_run": "2x:results/k76-capability-matrix-20261010-1059/2x/d0.json",
      "source_metric": "tg_total",
      "source_campaign": "k76-capability-matrix",
      "concurrency": 8,
      "input_tokens": 2048,
      "output_tokens": 512,
      "temperature": 1.0,
      "top_p": 0.95,
      "top_k": 20,
      "thinking": true,
      "cache_state": "prefix caching enabled; fresh 2048-token task, not a cold-cache campaign",
      "repetitions": "mean ± sd of 3 runs, one boot",
      "aggregate_denominator": "all output tokens / wall time including prompt processing",
      "per_request_denominator": "one output stream after first token",
      "release_image_digest": "sha256:8abb60eeeb79ca13680341c17d87975801da1218c06df2505856b23da7ed1821",
      "release_digest_is_proven_executed_digest": false
    },
    "evidence_state": "source_reported",
    "site_reproduced": false,
    "source_document_id": "qwen-tp2-csv",
    "source_document_sha256": "a82bffe4f0e33f4243db9549c02fb8e9379e5faa62825a35a7b087cd30b4996e",
    "source_line": 1158,
    "evidence_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv#L1158"
  },
  {
    "label": "TP2 v2.0.0 C16 2048 aggregate_output_rate",
    "metric": "aggregate_output_rate",
    "value": 231.294,
    "unit": "tokens/s",
    "statistic": "mean ± sd of 3 runs, one boot",
    "uncertainty": {
      "standard_deviation": 4.91,
      "kind": "SD across runs in one boot"
    },
    "conditions": {
      "model": "Qwen3.8-Flash-Next",
      "model_artifact": "ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE",
      "engine": "vLLM + b12x",
      "engine_version": null,
      "runtime_execution_revision": null,
      "executed_image_digest": null,
      "system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled",
      "node_count": 2,
      "parallelism": "TP2",
      "weight_precision": "NVFP4 routed experts and decode GDN; MXFP8 dense/attention and prefill GDN copies",
      "extra_prefill_revision": "7c4f1bc1a2d6847e0cbc01ac6b823f00251de8dd",
      "kv_dtype": "fp8",
      "draft_tokens": 4,
      "context_limit_tokens": 262144,
      "configured_max_requests": 16,
      "actual_output_tokens_per_request": null,
      "GPU_clock_cap_MHz": null,
      "model_revision": "16c9bd54788d12390838a65ce4a4ecda97fa5f1d",
      "recipe_release": "v2.0.0",
      "experiment_date": "2026-10-10",
      "harness": "llama-benchy 0.4.1.dev4+g0d4de4271",
      "source_run": "2x:results/k76-capability-matrix-20261010-1059/2x/d0.json",
      "source_metric": "tg_total",
      "source_campaign": "k76-capability-matrix",
      "concurrency": 16,
      "input_tokens": 2048,
      "output_tokens": 512,
      "temperature": 1.0,
      "top_p": 0.95,
      "top_k": 20,
      "thinking": true,
      "cache_state": "prefix caching enabled; fresh 2048-token task, not a cold-cache campaign",
      "repetitions": "mean ± sd of 3 runs, one boot",
      "aggregate_denominator": "all output tokens / wall time including prompt processing",
      "per_request_denominator": "one output stream after first token",
      "release_image_digest": "sha256:8abb60eeeb79ca13680341c17d87975801da1218c06df2505856b23da7ed1821",
      "release_digest_is_proven_executed_digest": false
    },
    "evidence_state": "source_reported",
    "site_reproduced": false,
    "source_document_id": "qwen-tp2-csv",
    "source_document_sha256": "a82bffe4f0e33f4243db9549c02fb8e9379e5faa62825a35a7b087cd30b4996e",
    "source_line": 1160,
    "evidence_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv#L1160"
  },
  {
    "label": "TP2 v2.2.0 C1 16384 ttft",
    "metric": "ttft",
    "value": 5.487,
    "unit": "s",
    "statistic": "mean of 4 boots x 3 cold prompts, sd between boots",
    "uncertainty": {
      "standard_deviation": 0.045,
      "kind": "SD between boot means; four boots x three cold prompts"
    },
    "conditions": {
      "model": "Qwen3.8-Flash-Next",
      "model_artifact": "ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE",
      "engine": "vLLM + b12x",
      "engine_version": null,
      "runtime_execution_revision": null,
      "executed_image_digest": null,
      "system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled",
      "node_count": 2,
      "parallelism": "TP2",
      "weight_precision": "NVFP4 routed experts and decode GDN; MXFP8 dense/attention and prefill GDN copies",
      "extra_prefill_revision": "7c4f1bc1a2d6847e0cbc01ac6b823f00251de8dd",
      "kv_dtype": "fp8",
      "draft_tokens": 4,
      "context_limit_tokens": 262144,
      "configured_max_requests": 16,
      "actual_output_tokens_per_request": null,
      "GPU_clock_cap_MHz": null,
      "model_revision": "03f4a0570496bbe741189f4a2a7d611679e85406",
      "recipe_release": "v2.2.0",
      "experiment_date": "2026-10-15",
      "harness": "cold.py (k101)",
      "source_run": "1x:results/k101s-1x-gdn-prefill-views-ab-20261015-1600/tp2/2x/",
      "source_metric": "ttft_s",
      "source_campaign": "promotion A/B",
      "concurrency": 1,
      "input_tokens": 16384,
      "output_tokens": 1,
      "temperature": 0,
      "top_p": null,
      "top_k": null,
      "thinking": null,
      "cache_state": "cold prefix with no hit, not a cold OS page-cache claim",
      "repetitions": "mean of 4 boots x 3 cold prompts, sd between boots",
      "aggregate_denominator": null,
      "per_request_denominator": null,
      "release_image_digest": null,
      "release_digest_is_proven_executed_digest": false
    },
    "evidence_state": "source_reported",
    "site_reproduced": false,
    "source_document_id": "qwen-tp2-csv",
    "source_document_sha256": "a82bffe4f0e33f4243db9549c02fb8e9379e5faa62825a35a7b087cd30b4996e",
    "source_line": 1262,
    "evidence_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv#L1262"
  },
  {
    "label": "TP2 v2.2.0 C1 16384 prefill_rate",
    "metric": "prefill_rate",
    "value": 2986.0,
    "unit": "tokens/s",
    "statistic": "mean of 4 boots x 3 cold prompts, sd between boots",
    "uncertainty": {
      "standard_deviation": 24.6,
      "kind": "SD between boot means; four boots x three cold prompts"
    },
    "conditions": {
      "model": "Qwen3.8-Flash-Next",
      "model_artifact": "ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE",
      "engine": "vLLM + b12x",
      "engine_version": null,
      "runtime_execution_revision": null,
      "executed_image_digest": null,
      "system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled",
      "node_count": 2,
      "parallelism": "TP2",
      "weight_precision": "NVFP4 routed experts and decode GDN; MXFP8 dense/attention and prefill GDN copies",
      "extra_prefill_revision": "7c4f1bc1a2d6847e0cbc01ac6b823f00251de8dd",
      "kv_dtype": "fp8",
      "draft_tokens": 4,
      "context_limit_tokens": 262144,
      "configured_max_requests": 16,
      "actual_output_tokens_per_request": null,
      "GPU_clock_cap_MHz": null,
      "model_revision": "03f4a0570496bbe741189f4a2a7d611679e85406",
      "recipe_release": "v2.2.0",
      "experiment_date": "2026-10-15",
      "harness": "cold.py (k101)",
      "source_run": "1x:results/k101s-1x-gdn-prefill-views-ab-20261015-1600/tp2/2x/",
      "source_metric": "pp",
      "source_campaign": "promotion A/B",
      "concurrency": 1,
      "input_tokens": 16384,
      "output_tokens": 1,
      "temperature": 0,
      "top_p": null,
      "top_k": null,
      "thinking": null,
      "cache_state": "cold prefix with no hit, not a cold OS page-cache claim",
      "repetitions": "mean of 4 boots x 3 cold prompts, sd between boots",
      "aggregate_denominator": null,
      "per_request_denominator": null,
      "release_image_digest": null,
      "release_digest_is_proven_executed_digest": false
    },
    "evidence_state": "source_reported",
    "site_reproduced": false,
    "source_document_id": "qwen-tp2-csv",
    "source_document_sha256": "a82bffe4f0e33f4243db9549c02fb8e9379e5faa62825a35a7b087cd30b4996e",
    "source_line": 1263,
    "evidence_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv#L1263"
  },
  {
    "label": "TP2 v2.2.0 C1 65536 ttft",
    "metric": "ttft",
    "value": 24.041,
    "unit": "s",
    "statistic": "mean of 4 boots x 3 cold prompts, sd between boots",
    "uncertainty": {
      "standard_deviation": 0.057,
      "kind": "SD between boot means; four boots x three cold prompts"
    },
    "conditions": {
      "model": "Qwen3.8-Flash-Next",
      "model_artifact": "ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE",
      "engine": "vLLM + b12x",
      "engine_version": null,
      "runtime_execution_revision": null,
      "executed_image_digest": null,
      "system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled",
      "node_count": 2,
      "parallelism": "TP2",
      "weight_precision": "NVFP4 routed experts and decode GDN; MXFP8 dense/attention and prefill GDN copies",
      "extra_prefill_revision": "7c4f1bc1a2d6847e0cbc01ac6b823f00251de8dd",
      "kv_dtype": "fp8",
      "draft_tokens": 4,
      "context_limit_tokens": 262144,
      "configured_max_requests": 16,
      "actual_output_tokens_per_request": null,
      "GPU_clock_cap_MHz": null,
      "model_revision": "03f4a0570496bbe741189f4a2a7d611679e85406",
      "recipe_release": "v2.2.0",
      "experiment_date": "2026-10-15",
      "harness": "cold.py (k101)",
      "source_run": "1x:results/k101s-1x-gdn-prefill-views-ab-20261015-1600/tp2/2x/",
      "source_metric": "ttft_s",
      "source_campaign": "promotion A/B",
      "concurrency": 1,
      "input_tokens": 65536,
      "output_tokens": 1,
      "temperature": 0,
      "top_p": null,
      "top_k": null,
      "thinking": null,
      "cache_state": "cold prefix with no hit, not a cold OS page-cache claim",
      "repetitions": "mean of 4 boots x 3 cold prompts, sd between boots",
      "aggregate_denominator": null,
      "per_request_denominator": null,
      "release_image_digest": null,
      "release_digest_is_proven_executed_digest": false
    },
    "evidence_state": "source_reported",
    "site_reproduced": false,
    "source_document_id": "qwen-tp2-csv",
    "source_document_sha256": "a82bffe4f0e33f4243db9549c02fb8e9379e5faa62825a35a7b087cd30b4996e",
    "source_line": 1264,
    "evidence_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv#L1264"
  },
  {
    "label": "TP2 v2.2.0 C1 65536 prefill_rate",
    "metric": "prefill_rate",
    "value": 2726.1,
    "unit": "tokens/s",
    "statistic": "mean of 4 boots x 3 cold prompts, sd between boots",
    "uncertainty": {
      "standard_deviation": 6.4,
      "kind": "SD between boot means; four boots x three cold prompts"
    },
    "conditions": {
      "model": "Qwen3.8-Flash-Next",
      "model_artifact": "ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE",
      "engine": "vLLM + b12x",
      "engine_version": null,
      "runtime_execution_revision": null,
      "executed_image_digest": null,
      "system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled",
      "node_count": 2,
      "parallelism": "TP2",
      "weight_precision": "NVFP4 routed experts and decode GDN; MXFP8 dense/attention and prefill GDN copies",
      "extra_prefill_revision": "7c4f1bc1a2d6847e0cbc01ac6b823f00251de8dd",
      "kv_dtype": "fp8",
      "draft_tokens": 4,
      "context_limit_tokens": 262144,
      "configured_max_requests": 16,
      "actual_output_tokens_per_request": null,
      "GPU_clock_cap_MHz": null,
      "model_revision": "03f4a0570496bbe741189f4a2a7d611679e85406",
      "recipe_release": "v2.2.0",
      "experiment_date": "2026-10-15",
      "harness": "cold.py (k101)",
      "source_run": "1x:results/k101s-1x-gdn-prefill-views-ab-20261015-1600/tp2/2x/",
      "source_metric": "pp",
      "source_campaign": "promotion A/B",
      "concurrency": 1,
      "input_tokens": 65536,
      "output_tokens": 1,
      "temperature": 0,
      "top_p": null,
      "top_k": null,
      "thinking": null,
      "cache_state": "cold prefix with no hit, not a cold OS page-cache claim",
      "repetitions": "mean of 4 boots x 3 cold prompts, sd between boots",
      "aggregate_denominator": null,
      "per_request_denominator": null,
      "release_image_digest": null,
      "release_digest_is_proven_executed_digest": false
    },
    "evidence_state": "source_reported",
    "site_reproduced": false,
    "source_document_id": "qwen-tp2-csv",
    "source_document_sha256": "a82bffe4f0e33f4243db9549c02fb8e9379e5faa62825a35a7b087cd30b4996e",
    "source_line": 1265,
    "evidence_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv#L1265"
  },
  {
    "label": "TP2 v2.2.0 C1 131072 ttft",
    "metric": "ttft",
    "value": 54.037,
    "unit": "s",
    "statistic": "mean of 4 boots x 3 cold prompts, sd between boots",
    "uncertainty": {
      "standard_deviation": 0.256,
      "kind": "SD between boot means; four boots x three cold prompts"
    },
    "conditions": {
      "model": "Qwen3.8-Flash-Next",
      "model_artifact": "ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE",
      "engine": "vLLM + b12x",
      "engine_version": null,
      "runtime_execution_revision": null,
      "executed_image_digest": null,
      "system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled",
      "node_count": 2,
      "parallelism": "TP2",
      "weight_precision": "NVFP4 routed experts and decode GDN; MXFP8 dense/attention and prefill GDN copies",
      "extra_prefill_revision": "7c4f1bc1a2d6847e0cbc01ac6b823f00251de8dd",
      "kv_dtype": "fp8",
      "draft_tokens": 4,
      "context_limit_tokens": 262144,
      "configured_max_requests": 16,
      "actual_output_tokens_per_request": null,
      "GPU_clock_cap_MHz": null,
      "model_revision": "03f4a0570496bbe741189f4a2a7d611679e85406",
      "recipe_release": "v2.2.0",
      "experiment_date": "2026-10-15",
      "harness": "cold.py (k101)",
      "source_run": "1x:results/k101s-1x-gdn-prefill-views-ab-20261015-1600/tp2/2x/",
      "source_metric": "ttft_s",
      "source_campaign": "promotion A/B",
      "concurrency": 1,
      "input_tokens": 131072,
      "output_tokens": 1,
      "temperature": 0,
      "top_p": null,
      "top_k": null,
      "thinking": null,
      "cache_state": "cold prefix with no hit, not a cold OS page-cache claim",
      "repetitions": "mean of 4 boots x 3 cold prompts, sd between boots",
      "aggregate_denominator": null,
      "per_request_denominator": null,
      "release_image_digest": null,
      "release_digest_is_proven_executed_digest": false
    },
    "evidence_state": "source_reported",
    "site_reproduced": false,
    "source_document_id": "qwen-tp2-csv",
    "source_document_sha256": "a82bffe4f0e33f4243db9549c02fb8e9379e5faa62825a35a7b087cd30b4996e",
    "source_line": 1266,
    "evidence_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv#L1266"
  },
  {
    "label": "TP2 v2.2.0 C1 131072 prefill_rate",
    "metric": "prefill_rate",
    "value": 2425.6,
    "unit": "tokens/s",
    "statistic": "mean of 4 boots x 3 cold prompts, sd between boots",
    "uncertainty": {
      "standard_deviation": 11.5,
      "kind": "SD between boot means; four boots x three cold prompts"
    },
    "conditions": {
      "model": "Qwen3.8-Flash-Next",
      "model_artifact": "ursuciprian/Qwen3.8-Flash-Next-NVFP4-GDN-MSE",
      "engine": "vLLM + b12x",
      "engine_version": null,
      "runtime_execution_revision": null,
      "executed_image_digest": null,
      "system": "two DGX Spark GB10, 128 GB unified memory each; ConnectX-7 direct-cabled",
      "node_count": 2,
      "parallelism": "TP2",
      "weight_precision": "NVFP4 routed experts and decode GDN; MXFP8 dense/attention and prefill GDN copies",
      "extra_prefill_revision": "7c4f1bc1a2d6847e0cbc01ac6b823f00251de8dd",
      "kv_dtype": "fp8",
      "draft_tokens": 4,
      "context_limit_tokens": 262144,
      "configured_max_requests": 16,
      "actual_output_tokens_per_request": null,
      "GPU_clock_cap_MHz": null,
      "model_revision": "03f4a0570496bbe741189f4a2a7d611679e85406",
      "recipe_release": "v2.2.0",
      "experiment_date": "2026-10-15",
      "harness": "cold.py (k101)",
      "source_run": "1x:results/k101s-1x-gdn-prefill-views-ab-20261015-1600/tp2/2x/",
      "source_metric": "pp",
      "source_campaign": "promotion A/B",
      "concurrency": 1,
      "input_tokens": 131072,
      "output_tokens": 1,
      "temperature": 0,
      "top_p": null,
      "top_k": null,
      "thinking": null,
      "cache_state": "cold prefix with no hit, not a cold OS page-cache claim",
      "repetitions": "mean of 4 boots x 3 cold prompts, sd between boots",
      "aggregate_denominator": null,
      "per_request_denominator": null,
      "release_image_digest": null,
      "release_digest_is_proven_executed_digest": false
    },
    "evidence_state": "source_reported",
    "site_reproduced": false,
    "source_document_id": "qwen-tp2-csv",
    "source_document_sha256": "a82bffe4f0e33f4243db9549c02fb8e9379e5faa62825a35a7b087cd30b4996e",
    "source_line": 1267,
    "evidence_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv#L1267"
  }
]
```

## Supplemental source identities

```json
{
  "supplemental_sources": [
    {
      "document_id": "qwen-tp2-csv",
      "parent_record_id": "qwen38-flash-next-tp2-concurrency-context",
      "publisher": "ursuciprian",
      "repository": "ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2",
      "revision": "b384f4a0e3ff0d563980befe52eb3e6964a5b6b1",
      "document_path": "docs/data/capability.csv",
      "pinned_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/data/capability.csv",
      "document_sha256": "a82bffe4f0e33f4243db9549c02fb8e9379e5faa62825a35a7b087cd30b4996e",
      "checked_on": "2026-10-18"
    },
    {
      "document_id": "qwen-tp2-method",
      "parent_record_id": "qwen38-flash-next-tp2-concurrency-context",
      "publisher": "ursuciprian",
      "repository": "ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2",
      "revision": "b384f4a0e3ff0d563980befe52eb3e6964a5b6b1",
      "document_path": "docs/BENCHMARKS.md",
      "pinned_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/docs/BENCHMARKS.md",
      "document_sha256": "5645ff4e129b3e2766672e092c65333b073b5550ad89aa6374cbd81ce1da7776",
      "checked_on": "2026-10-18"
    },
    {
      "document_id": "qwen-tp2-versions",
      "parent_record_id": "qwen38-flash-next-tp2-concurrency-context",
      "publisher": "ursuciprian",
      "repository": "ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2",
      "revision": "b384f4a0e3ff0d563980befe52eb3e6964a5b6b1",
      "document_path": "VERSIONS.md",
      "pinned_url": "https://github.com/ursuciprian/qwen3.8-flash-next-dgx-spark-tp-2/blob/b384f4a0e3ff0d563980befe52eb3e6964a5b6b1/VERSIONS.md",
      "document_sha256": "eb9ec6435802430a1be7302b445c67ceb48da9cf5e1ba32e939b5f7aaf9b100e",
      "checked_on": "2026-10-18"
    }
  ]
}
```