Benchmarks / Knowledge
Harvey LAB-AA v1.1
Harvey LAB-AA v1.1: 120 private agentic legal tasks built by Harvey. Hallucination-Gated All-Pass Rate: a task counts only when a three-judge panel passes every rubric criterion and the deliverables contain no material hallucination. Not comparable with v1.0 scores.
Harvey LAB-AA v1.1
Higher is better · best effort per model
full table below
| # | Model | Effort | Score |
|---|---|---|---|
| 1 | Muse Wonder Full (Max) | max | 13.1% |
| 2 | Muse Wonder | max | 10.0% |
| 3 | Grok 4.7 | xhigh | 9.4% |
| 4 | Muse Spark 1.3 | max | 8.9% |
| 5 | GPT-6 Astra | max | 8.6% |
| 6 | GPT-6.1 Sol | max | 6.9% |
| 7 | Fable 5.1 | max | 6.4% |
| 8 | Kimi K3 | max | 5.3% |
| 9 | Opus 5.5 | max | 4.2% |
| 10 | GPT-6 Sol | max | 3.6% |
| 11 | GPT-6 Luna | max | 3.3% |
| 12 | Mistral Large 4 | – | 3.1% |
| 13 | Sonnet 5.5 | max | 2.8% |
| 14 | Haiku 5.5 | xhigh | 2.8% |
| 15 | DeepSeek V4.1 Flash | max | 1.7% |
| 16 | GLM 5.3 Flash | max | 0.8% |
| 17 | GLM 5.3 | max | 0.3% |
| 18 | Gemini 3.8 Flash | high | 0.0% |
| 19 | Gemini 3.1 Pro | – | 0.0% |
| 20 | MiniMax M3 | – | 0.0% |
| 21 | Inkling Small | – | 0.0% |
| 22 | Inkling | xhigh | 0.0% |
| 23 | Gemini 3.5 Flash-Lite | high | 0.0% |