{
  "context": {
    "cap_tokens": 258816,
    "max_model_len": 262144,
    "targets": [
      {
        "attempted_context_tokens": 125000,
        "chars_per_token": 3.0,
        "clamped_tokens": 125000,
        "content_chunks": 3,
        "decode_tok_s": 133.7942077065515,
        "e2e_ms": 18598.38569999556,
        "effective_prefill_tok_s": 6578.79324147509,
        "estimated_prompt_tokens": 125000,
        "generation_ms": 4499.4473999977345,
        "mean_inter_token_latency_ms": 7.474165116275306,
        "output_token_source": "usage",
        "output_tokens": 603,
        "prompt_tokens": 92754,
        "reasoning_chunks": 172,
        "status": "passed",
        "target_tokens": 125000,
        "time_to_first_output_ms": 14098.938299997826,
        "ttft_ms": 18520.387500000652,
        "usage": {
          "completion_tokens": 603,
          "prompt_tokens": 92754,
          "prompt_tokens_details": {
            "cached_tokens": 0
          },
          "total_tokens": 93357
        },
        "visible_generation_ms": 77.99819999490865
      },
      {
        "attempted_context_tokens": 192000,
        "chars_per_token": 4.043,
        "clamped_tokens": 192000,
        "content_chunks": 10,
        "decode_tok_s": 126.50574745653279,
        "e2e_ms": 32135.741699996288,
        "effective_prefill_tok_s": 6274.787166803804,
        "estimated_prompt_tokens": 191999,
        "generation_ms": 1367.5267999933567,
        "mean_inter_token_latency_ms": 7.904779190713044,
        "output_token_source": "usage",
        "output_tokens": 174,
        "prompt_tokens": 193064,
        "reasoning_chunks": 45,
        "status": "passed",
        "target_tokens": 192000,
        "time_to_first_output_ms": 30768.21490000293,
        "ttft_ms": 31896.391999995103,
        "usage": {
          "completion_tokens": 174,
          "prompt_tokens": 193064,
          "prompt_tokens_details": {
            "cached_tokens": 4096
          },
          "total_tokens": 193238
        },
        "visible_generation_ms": 239.34970000118483
      },
      {
        "attempted_context_tokens": 250000,
        "chars_per_token": 4.021,
        "clamped_tokens": 250000,
        "content_chunks": 11,
        "decode_tok_s": 135.15138596995646,
        "e2e_ms": 45848.55140000582,
        "effective_prefill_tok_s": 5704.890292634713,
        "estimated_prompt_tokens": 249999,
        "generation_ms": 2101.3473000057274,
        "mean_inter_token_latency_ms": 7.3991102112877725,
        "output_token_source": "usage",
        "output_tokens": 285,
        "prompt_tokens": 249573,
        "reasoning_chunks": 72,
        "status": "passed",
        "target_tokens": 250000,
        "time_to_first_output_ms": 43747.20410000009,
        "ttft_ms": 45577.27760000853,
        "usage": {
          "completion_tokens": 285,
          "prompt_tokens": 249573,
          "prompt_tokens_details": {
            "cached_tokens": 5632
          },
          "total_tokens": 249858
        },
        "visible_generation_ms": 271.27379999728873
      }
    ]
  },
  "evaluation_protocol": {
    "budget_semantics": "visible_answer_tokens plus reasoning_headroom_tokens are sent as one max_tokens cap; the endpoint does not hard-partition the channels",
    "minimum_comparable_repetitions": 3,
    "minimum_pass_rate": 1.0,
    "reasoning_headroom_tokens": 1024,
    "records_finish_reason": true,
    "records_full_visible_answer": true,
    "records_reasoning_channel_metadata": true,
    "repetitions": 1,
    "version": 3,
    "visible_answer_tokens": 256
  },
  "failures": [],
  "identity": {
    "base_url": "http://127.0.0.1:39068/v1",
    "candidate_id": "deepseek-v4-flash-0731-r16-offload8-wsl2-mmap-unpinned",
    "config_id": "tp2-dspark5-256k",
    "engine": "vllm-r16-b12x-wsl2-offload-mmap-unpinned",
    "gpu": "2x RTX PRO 6000 Blackwell Max-Q TP2",
    "model": "deepseek-v4-flash-0731-r16-b12x-dspark5-offload8-wsl2-mmap-unpinned-tp2-256k",
    "started_at": "2026-08-02T05:24:01Z"
  },
  "intelligence": {
    "checks": [],
    "status": "not_run"
  },
  "run_id": "fast-bakeoff-20260802T052401Z",
  "schema": "anvil-serving.fast-tier-bakeoff/v1",
  "score_inputs": {
    "e2e_p50_ms": 32135.741699996288,
    "intelligence_pass_rate": null,
    "operational_fit_notes": [
      "endpoint was already loaded; benchmark did not start or stop serves"
    ],
    "session_recall_passed": false,
    "thinking_mode": "default",
    "tool_call_passed": false,
    "ttft_p50_ms": 31896.391999995103,
    "usable_context_tokens": 250000,
    "voice_latency_ms": null
  },
  "selection": {
    "context_targets": [
      125000,
      192000,
      250000
    ],
    "endpoint_already_loaded": true,
    "requests_per_context": 1,
    "suites": [
      "context"
    ]
  },
  "session": {
    "checks": [],
    "status": "not_run"
  },
  "source_recipe": {
    "ref": "configs/deepseek-v4-flash-0731-r16-b12x-dspark5-256k-offload8-wsl2-mmap-unpinned-recipe.toml",
    "serve_command": null
  },
  "suites": {},
  "thinking": {
    "chat_template_kwargs": null,
    "control_evidence": null,
    "control_evidence_sha256": null,
    "control_mechanism": "reasoning_effort",
    "control_requested": "low",
    "control_status": "requested_unverified",
    "mode": "default",
    "reasoning_effort": "low",
    "unsupported": false
  },
  "timing": {
    "chat": {
      "content_chunks": 24,
      "decode_tok_s_p50": 133.7942077065515,
      "decode_tok_s_p95": 135.15138596995646,
      "e2e_p50_ms": 32135.741699996288,
      "e2e_p95_ms": 45848.55140000582,
      "effective_prefill_tok_s_p50": 6274.787166803804,
      "effective_prefill_tok_s_p95": 6578.79324147509,
      "generation_p50_ms": 2101.3473000057274,
      "generation_p95_ms": 4499.4473999977345,
      "mean_inter_token_latency_ms_p50": 7.474165116275306,
      "mean_inter_token_latency_ms_p95": 7.904779190713044,
      "output_tokens": 1062,
      "prompt_token_samples": 3,
      "prompt_tokens": 535391,
      "prompt_tokens_p50": 193064.0,
      "prompt_tokens_p95": 249573.0,
      "reasoning_chunks": 289,
      "time_to_first_output_p50_ms": 30768.21490000293,
      "time_to_first_output_p95_ms": 43747.20410000009,
      "ttft_p50_ms": 31896.391999995103,
      "ttft_p95_ms": 45577.27760000853,
      "ttft_samples": 3
    },
    "wall_ms": 96591.90070000477
  },
  "tool": {
    "checks": [],
    "status": "not_run"
  },
  "voice": {
    "llm_latency_ms": null,
    "status": "not_run",
    "stt_latency_ms": null,
    "total_turn_latency_ms": null,
    "tts_latency_ms": null
  }
}
