Benchmark
Vals Harvey's Legal Agent Benchmark v1
Agentic benchmarks measure how well models carry out tasks that take several steps. Scores are percentages.
- Models measured
- 77
- Scores recorded
- 77
- Sources
- 1
- IQ weight
- 1.56
Score distribution
- 6.1%
- Mean
- 3.8%
- Median
- 6.3%
- Std dev
- 0%
- 10th percentile
- 15.8%
- 90th percentile
- 0%
- Min
This benchmark has a weight of 1.56 for chat models, measured over 77 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.
Leaderboard
| # | Model | Score |
|---|---|---|
| 51 | Kimi K2.6· Vals AI | 1.7% |
| 52 | MiMo V2.5· Vals AI | 1.7% |
| 53 | Qwen3.7 Max· Vals AI | 1.7% |
| 54 | Claude Haiku 5.5· Vals AI | 1.3% |
| 55 | GPT-5.6 Luna· Vals AI | 1.3% |
| 56 | Qwen3.6 Plus· Vals AI | 1.3% |
| 57 | Claude Haiku 4.5thinking· Vals AI | 0.833% |
| 58 | GPT-5.6 Terra· Vals AI | 0.833% |
| 59 | Grok 4.3· Vals AI | 0.417% |
| 60 | Mistral Medium 3.5· Vals AI | 0.417% |
| 61 | Nemotron 3 Ultra· Vals AI | 0.417% |
| 62 | Gemini 3 Flash· Vals AI | 0% |
| 63 | Gemini 3.1 Flash Lite· Vals AI | 0% |
| 64 | Gemini 3.1 Pro Preview· Vals AI | 0% |
| 65 | Gemini 3.5 Flash Lite· Vals AI | 0% |
| 66 | GLM 5.1· Vals AI | 0% |
| 67 | GPT 5.4· Vals AI | 0% |
| 68 | GPT 5.4 Nano· Vals AI | 0% |
| 69 | GPT-5.4 mini· Vals AI | 0% |
| 70 | Grok 4.20· Vals AI | 0% |
| 71 | Kimi K2.5thinking· Vals AI | 0% |
| 72 | Laguna M.1· Vals AI | 0% |
| 73 | Laguna XS.2· Vals AI | 0% |
| 74 | Mercury 2.5· Vals AI | 0% |
| 75 | MiniMax M2.7· Vals AI | 0% |
| 76 | Nemotron 3.5 Lightning 30B A3B· Vals AI | 0% |
| 77 | Qwen3.7 Plus· Vals AI | 0% |
Showing 51–77 of 77 results.