artificial-citizen commited on
Commit
de1ce89
·
verified ·
1 Parent(s): 969dcd7

Add vram_gb to all rows (artifact weight-file size) — unblocks /lab quality-vs-VRAM chart (protoLab#14, protoContent#394)

Browse files
README.md CHANGED
@@ -14,4 +14,8 @@ Methodology: single-stream-only numbers are never published without load numbers
14
  spec-decode accept% is reported per workload (random-data benches understate it ~2.5x);
15
  LLM-judged suites are gated behind deterministic ones. Harness: protoLabsAI/protoLab evals.
16
 
 
 
 
 
17
  CC-BY-4.0 — cite, steal, argue. Charts: protolabs.studio/lab
 
14
  spec-decode accept% is reported per workload (random-data benches understate it ~2.5x);
15
  LLM-judged suites are gated behind deterministic ones. Harness: protoLabsAI/protoLab evals.
16
 
17
+ `vram_gb` on each row is the artifact's published weight-file size in GB (the download you
18
+ load) — the traceable X-axis for the quality-vs-VRAM chart. GGUF rows use the specific quant
19
+ variant's file (e.g. the NVFP4 gguf, not a Q8_0 in the same repo), not the repo total.
20
+
21
  CC-BY-4.0 — cite, steal, argue. Charts: protolabs.studio/lab
agents-a1-2026-07-04.jsonl CHANGED
@@ -1,15 +1,15 @@
1
- {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "gate", "artifact": "protoLabsAI/Agents-A1-NVFP4", "quant": "nvfp4-w4a4-calibrated", "suite": "reasoning_v2", "n": 24, "metric": "mean_task_score", "value": 0.868, "baseline_fp8_official": 0.84, "grading": "solver"}
2
- {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "gate", "artifact": "protoLabsAI/Agents-A1-NVFP4", "quant": "nvfp4-w4a4-calibrated", "suite": "function_call", "n": 54, "metric": "pass_rate", "value": 0.907, "baseline_fp8_official": 0.889, "grading": "deterministic"}
3
- {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "gate", "artifact": "protoLabsAI/Agents-A1-NVFP4", "quant": "nvfp4-w4a4-calibrated", "suite": "claw_paired", "n": 85, "metric": "mean_task_score", "value": 0.648, "baseline_fp8_official": 0.64, "grading": "judge"}
4
- {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "gate", "artifact": "protoLabsAI/Agents-A1-NVFP4", "quant": "nvfp4-w4a4-calibrated", "suite": "code_spec_delta", "n": 8, "metric": "mean_task_score", "value": 0.453, "baseline_fp8_official": 0.476, "grading": "execution", "trials": 3}
5
- {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "coherence", "artifact": "protoLabsAI/Agents-A1-NVFP4", "quant": "nvfp4-w4a4-calibrated", "suite": "depth_probe", "value": "4/4 clean (4/16/32/60K; 60K thinking-off)", "grading": "detectors+adversarial-judge"}
6
- {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "speed", "artifact": "protoLabsAI/Agents-A1-NVFP4", "quant": "nvfp4-w4a4-calibrated", "config": "vllm-0.22.1 moe-backend=marlin", "hardware": "RTX PRO 6000 Blackwell sm120", "regime": "chat-1k1k", "concurrency": 1, "metric": "output_tok_s", "value": 215.2, "ttft_p50_ms": 63, "tpot_p50_ms": 4.6, "method": "vllm bench serve seeded random"}
7
- {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "speed", "artifact": "protoLabsAI/Agents-A1-NVFP4", "quant": "nvfp4-w4a4-calibrated", "config": "vllm-0.22.1 moe-backend=marlin", "hardware": "RTX PRO 6000 Blackwell sm120", "regime": "chat-1k1k", "concurrency": 8, "metric": "output_tok_s", "value": 1028.2, "ttft_p50_ms": 300, "tpot_p50_ms": 7.5, "method": "vllm bench serve seeded random"}
8
- {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "speed", "artifact": "protoLabsAI/Agents-A1-NVFP4", "quant": "nvfp4-w4a4-calibrated", "config": "vllm-0.22.1 moe-backend=marlin", "hardware": "RTX PRO 6000 Blackwell sm120", "regime": "context-8k1k", "concurrency": 1, "metric": "output_tok_s", "value": 200.1, "ttft_p50_ms": 313, "tpot_p50_ms": 4.7, "method": "vllm bench serve seeded random"}
9
- {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "speed", "artifact": "protoLabsAI/Agents-A1-NVFP4", "quant": "nvfp4-w4a4-calibrated", "config": "vllm-0.22.1 moe-backend=marlin", "hardware": "RTX PRO 6000 Blackwell sm120", "regime": "context-8k1k", "concurrency": 8, "metric": "output_tok_s", "value": 751.1, "ttft_p50_ms": 1630, "tpot_p50_ms": 9.1, "method": "vllm bench serve seeded random"}
10
- {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "gate", "artifact": "protoLabsAI/Agents-A1-MTP-GGUF", "quant": "nvfp4-gguf-mixed", "suite": "function_call", "metric": "pass_rate", "value": 0.889, "baseline_fp8_official": 0.889, "grading": "per-suite", "note": "identical to FP8"}
11
- {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "gate", "artifact": "protoLabsAI/Agents-A1-MTP-GGUF", "quant": "nvfp4-gguf-mixed", "suite": "reasoning_v2_16k", "metric": "mean_task_score", "value": 0.871, "baseline_fp8_official": 0.84, "grading": "per-suite", "note": "16K thinking budget; verbose-build caveat"}
12
- {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "gate", "artifact": "protoLabsAI/Agents-A1-MTP-GGUF", "quant": "nvfp4-gguf-mixed", "suite": "claw_paired", "metric": "mean_task_score", "value": 0.603, "baseline_fp8_official": 0.634, "grading": "per-suite", "note": "paired-84; zh cluster softer"}
13
- {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "speed", "artifact": "protoLabsAI/Agents-A1-MTP-GGUF", "quant": "nvfp4-gguf-mixed", "hardware": "RTX PRO 6000 Blackwell", "config": "llama.cpp draft-mtp", "metric": "gen_tok_s", "value": 305.0, "note": "mean n=6 prompts (287-336)"}
14
- {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "speed", "artifact": "protoLabsAI/Agents-A1-MTP-GGUF", "quant": "nvfp4-gguf-mixed", "hardware": "RTX A6000 Ampere", "config": "llama.cpp draft-mtp", "metric": "gen_tok_s", "value": 187.0, "note": "+46% over no-spec 127.7; accept 0.62"}
15
- {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "speed", "artifact": "protoLabsAI/Agents-A1-MTP-GGUF", "quant": "nvfp4-gguf-mixed", "hardware": "RTX A6000 Ampere", "config": "llama.cpp no-spec", "metric": "gen_tok_s", "value": 127.7, "note": ""}
 
1
+ {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "gate", "artifact": "protoLabsAI/Agents-A1-NVFP4", "quant": "nvfp4-w4a4-calibrated", "suite": "reasoning_v2", "n": 24, "metric": "mean_task_score", "value": 0.868, "baseline_fp8_official": 0.84, "grading": "solver", "vram_gb": 25.0}
2
+ {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "gate", "artifact": "protoLabsAI/Agents-A1-NVFP4", "quant": "nvfp4-w4a4-calibrated", "suite": "function_call", "n": 54, "metric": "pass_rate", "value": 0.907, "baseline_fp8_official": 0.889, "grading": "deterministic", "vram_gb": 25.0}
3
+ {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "gate", "artifact": "protoLabsAI/Agents-A1-NVFP4", "quant": "nvfp4-w4a4-calibrated", "suite": "claw_paired", "n": 85, "metric": "mean_task_score", "value": 0.648, "baseline_fp8_official": 0.64, "grading": "judge", "vram_gb": 25.0}
4
+ {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "gate", "artifact": "protoLabsAI/Agents-A1-NVFP4", "quant": "nvfp4-w4a4-calibrated", "suite": "code_spec_delta", "n": 8, "metric": "mean_task_score", "value": 0.453, "baseline_fp8_official": 0.476, "grading": "execution", "trials": 3, "vram_gb": 25.0}
5
+ {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "coherence", "artifact": "protoLabsAI/Agents-A1-NVFP4", "quant": "nvfp4-w4a4-calibrated", "suite": "depth_probe", "value": "4/4 clean (4/16/32/60K; 60K thinking-off)", "grading": "detectors+adversarial-judge", "vram_gb": 25.0}
6
+ {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "speed", "artifact": "protoLabsAI/Agents-A1-NVFP4", "quant": "nvfp4-w4a4-calibrated", "config": "vllm-0.22.1 moe-backend=marlin", "hardware": "RTX PRO 6000 Blackwell sm120", "regime": "chat-1k1k", "concurrency": 1, "metric": "output_tok_s", "value": 215.2, "ttft_p50_ms": 63, "tpot_p50_ms": 4.6, "method": "vllm bench serve seeded random", "vram_gb": 25.0}
7
+ {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "speed", "artifact": "protoLabsAI/Agents-A1-NVFP4", "quant": "nvfp4-w4a4-calibrated", "config": "vllm-0.22.1 moe-backend=marlin", "hardware": "RTX PRO 6000 Blackwell sm120", "regime": "chat-1k1k", "concurrency": 8, "metric": "output_tok_s", "value": 1028.2, "ttft_p50_ms": 300, "tpot_p50_ms": 7.5, "method": "vllm bench serve seeded random", "vram_gb": 25.0}
8
+ {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "speed", "artifact": "protoLabsAI/Agents-A1-NVFP4", "quant": "nvfp4-w4a4-calibrated", "config": "vllm-0.22.1 moe-backend=marlin", "hardware": "RTX PRO 6000 Blackwell sm120", "regime": "context-8k1k", "concurrency": 1, "metric": "output_tok_s", "value": 200.1, "ttft_p50_ms": 313, "tpot_p50_ms": 4.7, "method": "vllm bench serve seeded random", "vram_gb": 25.0}
9
+ {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "speed", "artifact": "protoLabsAI/Agents-A1-NVFP4", "quant": "nvfp4-w4a4-calibrated", "config": "vllm-0.22.1 moe-backend=marlin", "hardware": "RTX PRO 6000 Blackwell sm120", "regime": "context-8k1k", "concurrency": 8, "metric": "output_tok_s", "value": 751.1, "ttft_p50_ms": 1630, "tpot_p50_ms": 9.1, "method": "vllm bench serve seeded random", "vram_gb": 25.0}
10
+ {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "gate", "artifact": "protoLabsAI/Agents-A1-MTP-GGUF", "quant": "nvfp4-gguf-mixed", "suite": "function_call", "metric": "pass_rate", "value": 0.889, "baseline_fp8_official": 0.889, "grading": "per-suite", "note": "identical to FP8", "vram_gb": 20.8}
11
+ {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "gate", "artifact": "protoLabsAI/Agents-A1-MTP-GGUF", "quant": "nvfp4-gguf-mixed", "suite": "reasoning_v2_16k", "metric": "mean_task_score", "value": 0.871, "baseline_fp8_official": 0.84, "grading": "per-suite", "note": "16K thinking budget; verbose-build caveat", "vram_gb": 20.8}
12
+ {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "gate", "artifact": "protoLabsAI/Agents-A1-MTP-GGUF", "quant": "nvfp4-gguf-mixed", "suite": "claw_paired", "metric": "mean_task_score", "value": 0.603, "baseline_fp8_official": 0.634, "grading": "per-suite", "note": "paired-84; zh cluster softer", "vram_gb": 20.8}
13
+ {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "speed", "artifact": "protoLabsAI/Agents-A1-MTP-GGUF", "quant": "nvfp4-gguf-mixed", "hardware": "RTX PRO 6000 Blackwell", "config": "llama.cpp draft-mtp", "metric": "gen_tok_s", "value": 305.0, "note": "mean n=6 prompts (287-336)", "vram_gb": 20.8}
14
+ {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "speed", "artifact": "protoLabsAI/Agents-A1-MTP-GGUF", "quant": "nvfp4-gguf-mixed", "hardware": "RTX A6000 Ampere", "config": "llama.cpp draft-mtp", "metric": "gen_tok_s", "value": 187.0, "note": "+46% over no-spec 127.7; accept 0.62", "vram_gb": 20.8}
15
+ {"model": "InternScience/Agents-A1", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "speed", "artifact": "protoLabsAI/Agents-A1-MTP-GGUF", "quant": "nvfp4-gguf-mixed", "hardware": "RTX A6000 Ampere", "config": "llama.cpp no-spec", "metric": "gen_tok_s", "value": 127.7, "note": "", "vram_gb": 20.8}
ornith-35b-nvfp4-2026-07-04.jsonl CHANGED
@@ -1,9 +1,9 @@
1
- {"model": "deepreinforce-ai/Ornith-1.0-35B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-35B-NVFP4", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "gate", "suite": "claw_paired", "n": 35, "metric": "mean_task_score", "value": 0.677, "baseline_fp8_prod": 0.672, "trials": 1, "grading": "judged", "note": "T28 excluded; 3 outliers re-trialed x3 both sides (T26 flip was baseline drift, not quant)"}
2
- {"model": "deepreinforce-ai/Ornith-1.0-35B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-35B-NVFP4", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "gate", "suite": "function_call", "n": 54, "metric": "pass_rate", "value": 0.944, "baseline_fp8_prod": 0.93, "trials": 1, "grading": "deterministic"}
3
- {"model": "deepreinforce-ai/Ornith-1.0-35B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-35B-NVFP4", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "gate", "suite": "reasoning_v2", "n": 24, "metric": "mean_task_score", "value": 0.889, "stddev": 0.042, "baseline_fp8_prod": 0.882, "baseline_stddev": 0.021, "trials": 3, "baseline_trials": 3, "grading": "solver"}
4
- {"model": "deepreinforce-ai/Ornith-1.0-35B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-35B-NVFP4", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "gate", "suite": "code_spec_delta", "n": 8, "metric": "mean_score", "value": 0.58, "stddev": 0.062, "baseline_fp8_prod": 0.616, "trials": 3, "grading": "exec"}
5
- {"model": "deepreinforce-ai/Ornith-1.0-35B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-35B-NVFP4", "date": "2026-07-04", "harness": "verify_coherence adversarial", "judge": "protolabs/reasoning", "kind": "coherence", "suite": "depth_coherence", "config": "4096,16384,32768,60000", "metric": "rungs_clean", "value": 4, "n": 4, "grading": "detectors+judge", "note": "needle recall perfect all rungs, zero degeneration flags"}
6
- {"model": "deepreinforce-ai/Ornith-1.0-35B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-35B-NVFP4", "date": "2026-07-04", "kind": "speed", "engine": "vLLM 0.22.1", "hardware": "RTX PRO 6000 Blackwell (sm120), 1x GPU", "method": "speed-test-v2 quick (vllm bench serve, random, fixed seed)", "regime": "chat_1k_1k", "concurrency": 1, "metric": "agg_tok_s", "value": 207.8, "ttft_p50_ms": 62, "tpot_p50_ms": 4.8, "goodput": 0.2}
7
- {"model": "deepreinforce-ai/Ornith-1.0-35B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-35B-NVFP4", "date": "2026-07-04", "kind": "speed", "engine": "vLLM 0.22.1", "hardware": "RTX PRO 6000 Blackwell (sm120), 1x GPU", "method": "speed-test-v2 quick (vllm bench serve, random, fixed seed)", "regime": "chat_1k_1k", "concurrency": 8, "metric": "agg_tok_s", "value": 950.6, "ttft_p50_ms": 297, "tpot_p50_ms": 7.8, "goodput": 0.81}
8
- {"model": "deepreinforce-ai/Ornith-1.0-35B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-35B-NVFP4", "date": "2026-07-04", "kind": "speed", "engine": "vLLM 0.22.1", "hardware": "RTX PRO 6000 Blackwell (sm120), 1x GPU", "method": "speed-test-v2 quick (vllm bench serve, random, fixed seed)", "regime": "context_8k_1k", "concurrency": 1, "metric": "agg_tok_s", "value": 194.2, "ttft_p50_ms": 305, "tpot_p50_ms": 4.9, "goodput": 0.19}
9
- {"model": "deepreinforce-ai/Ornith-1.0-35B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-35B-NVFP4", "date": "2026-07-04", "kind": "speed", "engine": "vLLM 0.22.1", "hardware": "RTX PRO 6000 Blackwell (sm120), 1x GPU", "method": "speed-test-v2 quick (vllm bench serve, random, fixed seed)", "regime": "context_8k_1k", "concurrency": 8, "metric": "agg_tok_s", "value": 738.4, "ttft_p50_ms": 1520, "tpot_p50_ms": 9.3, "goodput": 0.64}
 
1
+ {"model": "deepreinforce-ai/Ornith-1.0-35B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-35B-NVFP4", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "gate", "suite": "claw_paired", "n": 35, "metric": "mean_task_score", "value": 0.677, "baseline_fp8_prod": 0.672, "trials": 1, "grading": "judged", "note": "T28 excluded; 3 outliers re-trialed x3 both sides (T26 flip was baseline drift, not quant)", "vram_gb": 25.0}
2
+ {"model": "deepreinforce-ai/Ornith-1.0-35B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-35B-NVFP4", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "gate", "suite": "function_call", "n": 54, "metric": "pass_rate", "value": 0.944, "baseline_fp8_prod": 0.93, "trials": 1, "grading": "deterministic", "vram_gb": 25.0}
3
+ {"model": "deepreinforce-ai/Ornith-1.0-35B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-35B-NVFP4", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "gate", "suite": "reasoning_v2", "n": 24, "metric": "mean_task_score", "value": 0.889, "stddev": 0.042, "baseline_fp8_prod": 0.882, "baseline_stddev": 0.021, "trials": 3, "baseline_trials": 3, "grading": "solver", "vram_gb": 25.0}
4
+ {"model": "deepreinforce-ai/Ornith-1.0-35B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-35B-NVFP4", "date": "2026-07-04", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "budget": 8192, "kind": "gate", "suite": "code_spec_delta", "n": 8, "metric": "mean_score", "value": 0.58, "stddev": 0.062, "baseline_fp8_prod": 0.616, "trials": 3, "grading": "exec", "vram_gb": 25.0}
5
+ {"model": "deepreinforce-ai/Ornith-1.0-35B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-35B-NVFP4", "date": "2026-07-04", "harness": "verify_coherence adversarial", "judge": "protolabs/reasoning", "kind": "coherence", "suite": "depth_coherence", "config": "4096,16384,32768,60000", "metric": "rungs_clean", "value": 4, "n": 4, "grading": "detectors+judge", "note": "needle recall perfect all rungs, zero degeneration flags", "vram_gb": 25.0}
6
+ {"model": "deepreinforce-ai/Ornith-1.0-35B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-35B-NVFP4", "date": "2026-07-04", "kind": "speed", "engine": "vLLM 0.22.1", "hardware": "RTX PRO 6000 Blackwell (sm120), 1x GPU", "method": "speed-test-v2 quick (vllm bench serve, random, fixed seed)", "regime": "chat_1k_1k", "concurrency": 1, "metric": "agg_tok_s", "value": 207.8, "ttft_p50_ms": 62, "tpot_p50_ms": 4.8, "goodput": 0.2, "vram_gb": 25.0}
7
+ {"model": "deepreinforce-ai/Ornith-1.0-35B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-35B-NVFP4", "date": "2026-07-04", "kind": "speed", "engine": "vLLM 0.22.1", "hardware": "RTX PRO 6000 Blackwell (sm120), 1x GPU", "method": "speed-test-v2 quick (vllm bench serve, random, fixed seed)", "regime": "chat_1k_1k", "concurrency": 8, "metric": "agg_tok_s", "value": 950.6, "ttft_p50_ms": 297, "tpot_p50_ms": 7.8, "goodput": 0.81, "vram_gb": 25.0}
8
+ {"model": "deepreinforce-ai/Ornith-1.0-35B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-35B-NVFP4", "date": "2026-07-04", "kind": "speed", "engine": "vLLM 0.22.1", "hardware": "RTX PRO 6000 Blackwell (sm120), 1x GPU", "method": "speed-test-v2 quick (vllm bench serve, random, fixed seed)", "regime": "context_8k_1k", "concurrency": 1, "metric": "agg_tok_s", "value": 194.2, "ttft_p50_ms": 305, "tpot_p50_ms": 4.9, "goodput": 0.19, "vram_gb": 25.0}
9
+ {"model": "deepreinforce-ai/Ornith-1.0-35B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-35B-NVFP4", "date": "2026-07-04", "kind": "speed", "engine": "vLLM 0.22.1", "hardware": "RTX PRO 6000 Blackwell (sm120), 1x GPU", "method": "speed-test-v2 quick (vllm bench serve, random, fixed seed)", "regime": "context_8k_1k", "concurrency": 8, "metric": "agg_tok_s", "value": 738.4, "ttft_p50_ms": 1520, "tpot_p50_ms": 9.3, "goodput": 0.64, "vram_gb": 25.0}
ornith-9b-nvfp4-2026-07-03.jsonl CHANGED
@@ -1,29 +1,29 @@
1
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "gate", "suite": "function_call", "n": 54, "metric": "pass_rate", "value": 0.96, "baseline_bf16": 0.93, "trials": 1, "grading": "deterministic"}
2
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "gate", "suite": "reasoning_v2", "n": 24, "metric": "mean_task_score", "value": 0.684, "baseline_bf16": 0.726, "trials": 1, "grading": "solver"}
3
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "gate", "suite": "code_exec_v2", "n": 8, "metric": "mean_task_score", "value": 0.405, "baseline_bf16": 0.391, "trials": 3, "grading": "execution"}
4
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "gate", "suite": "claw_paired29", "n": 29, "metric": "mean_task_score", "value": 0.791, "baseline_bf16": 0.819, "trials": "1 (+3x outliers both sides)", "grading": "judge", "note": "T28 excluded (bimodal judge); T12 reproducible miss 0/3 (duplicate detection)"}
5
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "coherence", "suite": "depth_probe", "n": 4, "metric": "rungs_clean", "value": "4/4 (4K/16K/32K/60K)", "grading": "detectors+adversarial-judge"}
6
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "nvfp4+mtp", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "chat-c1", "metric": "output_tok_s", "value": 132.2, "ttft_p50_ms": 46.8, "tpot_p50_ms": 7.69, "method": "vllm bench serve, random dataset seed 42, client-side"}
7
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "nvfp4+mtp", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "chat-c8", "metric": "output_tok_s", "value": 907.1, "ttft_p50_ms": 70.5, "tpot_p50_ms": 8.01, "method": "vllm bench serve, random dataset seed 42, client-side"}
8
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "nvfp4+mtp", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "context-c1", "metric": "output_tok_s", "value": 108.1, "ttft_p50_ms": 343.2, "tpot_p50_ms": 8.92, "method": "vllm bench serve, random dataset seed 42, client-side"}
9
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "nvfp4+mtp", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "context-c8", "metric": "output_tok_s", "value": 598.0, "ttft_p50_ms": 412.3, "tpot_p50_ms": 12.23, "method": "vllm bench serve, random dataset seed 42, client-side"}
10
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "bf16+mtp", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "chat-c1", "metric": "output_tok_s", "value": 82.7, "ttft_p50_ms": 90.1, "tpot_p50_ms": 12.4, "method": "vllm bench serve, random dataset seed 42, client-side"}
11
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "bf16+mtp", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "chat-c8", "metric": "output_tok_s", "value": 620.8, "ttft_p50_ms": 121.0, "tpot_p50_ms": 12.37, "method": "vllm bench serve, random dataset seed 42, client-side"}
12
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "bf16+mtp", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "context-c1", "metric": "output_tok_s", "value": 73.1, "ttft_p50_ms": 656.8, "tpot_p50_ms": 13.24, "method": "vllm bench serve, random dataset seed 42, client-side"}
13
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "bf16+mtp", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "context-c8", "metric": "output_tok_s", "value": 385.9, "ttft_p50_ms": 733.3, "tpot_p50_ms": 18.45, "method": "vllm bench serve, random dataset seed 42, client-side"}
14
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "nvfp4+mtp-depth", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "d16k-c1", "metric": "output_tok_s", "value": 64.1, "ttft_p50_ms": 732.0, "tpot_p50_ms": 13.28, "method": "vllm bench serve, random dataset seed 42, client-side"}
15
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "nvfp4+mtp-depth", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "d16k-c4", "metric": "output_tok_s", "value": 159.7, "ttft_p50_ms": 834.8, "tpot_p50_ms": 21.64, "method": "vllm bench serve, random dataset seed 42, client-side"}
16
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "nvfp4+mtp-depth", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "d32k-c1", "metric": "output_tok_s", "value": 42.1, "ttft_p50_ms": 1709.3, "tpot_p50_ms": 17.92, "method": "vllm bench serve, random dataset seed 42, client-side"}
17
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "nvfp4+mtp-depth", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "d32k-c4", "metric": "output_tok_s", "value": 88.9, "ttft_p50_ms": 2265.5, "tpot_p50_ms": 36.8, "method": "vllm bench serve, random dataset seed 42, client-side"}
18
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "nvfp4+mtp-depth", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "d4k-c1", "metric": "output_tok_s", "value": 104.3, "ttft_p50_ms": 139.2, "tpot_p50_ms": 9.5, "method": "vllm bench serve, random dataset seed 42, client-side"}
19
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "nvfp4+mtp-depth", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "d4k-c4", "metric": "output_tok_s", "value": 311.1, "ttft_p50_ms": 191.7, "tpot_p50_ms": 11.38, "method": "vllm bench serve, random dataset seed 42, client-side"}
20
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "nvfp4+mtp-depth", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "d64k-c1", "metric": "output_tok_s", "value": 24.3, "ttft_p50_ms": 4081.7, "tpot_p50_ms": 26.9, "method": "vllm bench serve, random dataset seed 42, client-side"}
21
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "nvfp4+mtp-depth", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "d64k-c4", "metric": "output_tok_s", "value": 44.4, "ttft_p50_ms": 4631.7, "tpot_p50_ms": 71.76, "method": "vllm bench serve, random dataset seed 42, client-side"}
22
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "gguf-Q4_K_M-spec-none", "hardware": "ampere-a6000", "engine": "llama.cpp b9829", "metric": "gen_tok_s", "value": 104.6, "method": "llama-cli -n 200 greedy; blackwell mtp = mean n=6 prompts", "note": ""}
23
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "gguf-Q4_K_M-spec-draft-mtp", "hardware": "ampere-a6000", "engine": "llama.cpp b9829", "metric": "gen_tok_s", "value": 153.4, "method": "llama-cli -n 200 greedy; blackwell mtp = mean n=6 prompts", "note": "acceptance 0.52 (NVFP4) vs 0.49 (Q4_K_M) same-box"}
24
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "gguf-NVFP4-spec-none", "hardware": "ampere-a6000", "engine": "llama.cpp b9829", "metric": "gen_tok_s", "value": 70.7, "method": "llama-cli -n 200 greedy; blackwell mtp = mean n=6 prompts", "note": ""}
25
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "gguf-NVFP4-spec-draft-mtp", "hardware": "ampere-a6000", "engine": "llama.cpp b9829", "metric": "gen_tok_s", "value": 84.8, "method": "llama-cli -n 200 greedy; blackwell mtp = mean n=6 prompts", "note": "acceptance 0.52 (NVFP4) vs 0.49 (Q4_K_M) same-box"}
26
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "gguf-Q4_K_M-spec-none", "hardware": "blackwell-sm120", "engine": "llama.cpp b9829", "metric": "gen_tok_s", "value": 205.1, "method": "llama-cli -n 200 greedy; blackwell mtp = mean n=6 prompts", "note": ""}
27
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "gguf-Q4_K_M-spec-draft-mtp", "hardware": "blackwell-sm120", "engine": "llama.cpp b9829", "metric": "gen_tok_s", "value": 239.0, "method": "llama-cli -n 200 greedy; blackwell mtp = mean n=6 prompts", "note": "acceptance 0.52 (NVFP4) vs 0.49 (Q4_K_M) same-box"}
28
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "gguf-NVFP4-spec-none", "hardware": "blackwell-sm120", "engine": "llama.cpp b9829", "metric": "gen_tok_s", "value": 201.5, "method": "llama-cli -n 200 greedy; blackwell mtp = mean n=6 prompts", "note": ""}
29
- {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "gguf-NVFP4-spec-draft-mtp", "hardware": "blackwell-sm120", "engine": "llama.cpp b9829", "metric": "gen_tok_s", "value": 306.0, "method": "llama-cli -n 200 greedy; blackwell mtp = mean n=6 prompts", "note": "acceptance 0.52 (NVFP4) vs 0.49 (Q4_K_M) same-box"}
 
1
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "gate", "suite": "function_call", "n": 54, "metric": "pass_rate", "value": 0.96, "baseline_bf16": 0.93, "trials": 1, "grading": "deterministic", "vram_gb": 11.7}
2
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "gate", "suite": "reasoning_v2", "n": 24, "metric": "mean_task_score", "value": 0.684, "baseline_bf16": 0.726, "trials": 1, "grading": "solver", "vram_gb": 11.7}
3
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "gate", "suite": "code_exec_v2", "n": 8, "metric": "mean_task_score", "value": 0.405, "baseline_bf16": 0.391, "trials": 3, "grading": "execution", "vram_gb": 11.7}
4
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "gate", "suite": "claw_paired29", "n": 29, "metric": "mean_task_score", "value": 0.791, "baseline_bf16": 0.819, "trials": "1 (+3x outliers both sides)", "grading": "judge", "note": "T28 excluded (bimodal judge); T12 reproducible miss 0/3 (duplicate detection)", "vram_gb": 11.7}
5
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "coherence", "suite": "depth_probe", "n": 4, "metric": "rungs_clean", "value": "4/4 (4K/16K/32K/60K)", "grading": "detectors+adversarial-judge", "vram_gb": 11.7}
6
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "nvfp4+mtp", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "chat-c1", "metric": "output_tok_s", "value": 132.2, "ttft_p50_ms": 46.8, "tpot_p50_ms": 7.69, "method": "vllm bench serve, random dataset seed 42, client-side", "vram_gb": 11.7}
7
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "nvfp4+mtp", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "chat-c8", "metric": "output_tok_s", "value": 907.1, "ttft_p50_ms": 70.5, "tpot_p50_ms": 8.01, "method": "vllm bench serve, random dataset seed 42, client-side", "vram_gb": 11.7}
8
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "nvfp4+mtp", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "context-c1", "metric": "output_tok_s", "value": 108.1, "ttft_p50_ms": 343.2, "tpot_p50_ms": 8.92, "method": "vllm bench serve, random dataset seed 42, client-side", "vram_gb": 11.7}
9
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "nvfp4+mtp", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "context-c8", "metric": "output_tok_s", "value": 598.0, "ttft_p50_ms": 412.3, "tpot_p50_ms": 12.23, "method": "vllm bench serve, random dataset seed 42, client-side", "vram_gb": 11.7}
10
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "bf16+mtp", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "chat-c1", "metric": "output_tok_s", "value": 82.7, "ttft_p50_ms": 90.1, "tpot_p50_ms": 12.4, "method": "vllm bench serve, random dataset seed 42, client-side", "vram_gb": 11.7}
11
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "bf16+mtp", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "chat-c8", "metric": "output_tok_s", "value": 620.8, "ttft_p50_ms": 121.0, "tpot_p50_ms": 12.37, "method": "vllm bench serve, random dataset seed 42, client-side", "vram_gb": 11.7}
12
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "bf16+mtp", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "context-c1", "metric": "output_tok_s", "value": 73.1, "ttft_p50_ms": 656.8, "tpot_p50_ms": 13.24, "method": "vllm bench serve, random dataset seed 42, client-side", "vram_gb": 11.7}
13
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "bf16+mtp", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "context-c8", "metric": "output_tok_s", "value": 385.9, "ttft_p50_ms": 733.3, "tpot_p50_ms": 18.45, "method": "vllm bench serve, random dataset seed 42, client-side", "vram_gb": 11.7}
14
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "nvfp4+mtp-depth", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "d16k-c1", "metric": "output_tok_s", "value": 64.1, "ttft_p50_ms": 732.0, "tpot_p50_ms": 13.28, "method": "vllm bench serve, random dataset seed 42, client-side", "vram_gb": 11.7}
15
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "nvfp4+mtp-depth", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "d16k-c4", "metric": "output_tok_s", "value": 159.7, "ttft_p50_ms": 834.8, "tpot_p50_ms": 21.64, "method": "vllm bench serve, random dataset seed 42, client-side", "vram_gb": 11.7}
16
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "nvfp4+mtp-depth", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "d32k-c1", "metric": "output_tok_s", "value": 42.1, "ttft_p50_ms": 1709.3, "tpot_p50_ms": 17.92, "method": "vllm bench serve, random dataset seed 42, client-side", "vram_gb": 11.7}
17
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "nvfp4+mtp-depth", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "d32k-c4", "metric": "output_tok_s", "value": 88.9, "ttft_p50_ms": 2265.5, "tpot_p50_ms": 36.8, "method": "vllm bench serve, random dataset seed 42, client-side", "vram_gb": 11.7}
18
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "nvfp4+mtp-depth", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "d4k-c1", "metric": "output_tok_s", "value": 104.3, "ttft_p50_ms": 139.2, "tpot_p50_ms": 9.5, "method": "vllm bench serve, random dataset seed 42, client-side", "vram_gb": 11.7}
19
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "nvfp4+mtp-depth", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "d4k-c4", "metric": "output_tok_s", "value": 311.1, "ttft_p50_ms": 191.7, "tpot_p50_ms": 11.38, "method": "vllm bench serve, random dataset seed 42, client-side", "vram_gb": 11.7}
20
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "nvfp4+mtp-depth", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "d64k-c1", "metric": "output_tok_s", "value": 24.3, "ttft_p50_ms": 4081.7, "tpot_p50_ms": 26.9, "method": "vllm bench serve, random dataset seed 42, client-side", "vram_gb": 11.7}
21
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "nvfp4+mtp-depth", "hardware": "RTX PRO 6000 Blackwell (sm120)", "engine": "vLLM 0.22.1", "regime": "d64k-c4", "metric": "output_tok_s", "value": 44.4, "ttft_p50_ms": 4631.7, "tpot_p50_ms": 71.76, "method": "vllm bench serve, random dataset seed 42, client-side", "vram_gb": 11.7}
22
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "gguf-Q4_K_M-spec-none", "hardware": "ampere-a6000", "engine": "llama.cpp b9829", "metric": "gen_tok_s", "value": 104.6, "method": "llama-cli -n 200 greedy; blackwell mtp = mean n=6 prompts", "note": "", "vram_gb": 11.7}
23
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "gguf-Q4_K_M-spec-draft-mtp", "hardware": "ampere-a6000", "engine": "llama.cpp b9829", "metric": "gen_tok_s", "value": 153.4, "method": "llama-cli -n 200 greedy; blackwell mtp = mean n=6 prompts", "note": "acceptance 0.52 (NVFP4) vs 0.49 (Q4_K_M) same-box", "vram_gb": 11.7}
24
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "gguf-NVFP4-spec-none", "hardware": "ampere-a6000", "engine": "llama.cpp b9829", "metric": "gen_tok_s", "value": 70.7, "method": "llama-cli -n 200 greedy; blackwell mtp = mean n=6 prompts", "note": "", "vram_gb": 11.7}
25
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "gguf-NVFP4-spec-draft-mtp", "hardware": "ampere-a6000", "engine": "llama.cpp b9829", "metric": "gen_tok_s", "value": 84.8, "method": "llama-cli -n 200 greedy; blackwell mtp = mean n=6 prompts", "note": "acceptance 0.52 (NVFP4) vs 0.49 (Q4_K_M) same-box", "vram_gb": 11.7}
26
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "gguf-Q4_K_M-spec-none", "hardware": "blackwell-sm120", "engine": "llama.cpp b9829", "metric": "gen_tok_s", "value": 205.1, "method": "llama-cli -n 200 greedy; blackwell mtp = mean n=6 prompts", "note": "", "vram_gb": 11.7}
27
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "gguf-Q4_K_M-spec-draft-mtp", "hardware": "blackwell-sm120", "engine": "llama.cpp b9829", "metric": "gen_tok_s", "value": 239.0, "method": "llama-cli -n 200 greedy; blackwell mtp = mean n=6 prompts", "note": "acceptance 0.52 (NVFP4) vs 0.49 (Q4_K_M) same-box", "vram_gb": 11.7}
28
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "gguf-NVFP4-spec-none", "hardware": "blackwell-sm120", "engine": "llama.cpp b9829", "metric": "gen_tok_s", "value": 201.5, "method": "llama-cli -n 200 greedy; blackwell mtp = mean n=6 prompts", "note": "", "vram_gb": 11.7}
29
+ {"model": "deepreinforce-ai/Ornith-1.0-9B", "quant": "nvfp4-w4a4-calibrated", "artifact": "protoLabsAI/Ornith-1.0-9B-NVFP4", "date": "2026-07-03", "harness": "protoLab evals @ claw-eval 49d2b6a", "judge": "protolabs/reasoning", "kind": "speed", "config": "gguf-NVFP4-spec-draft-mtp", "hardware": "blackwell-sm120", "engine": "llama.cpp b9829", "metric": "gen_tok_s", "value": 306.0, "method": "llama-cli -n 200 greedy; blackwell mtp = mean n=6 prompts", "note": "acceptance 0.52 (NVFP4) vs 0.49 (Q4_K_M) same-box", "vram_gb": 11.7}