Benchmarks · knowledge
CaseLaw (Vals)
Private question-answer benchmark over Canadian court-cases. Scores are the values published by VALS AI. Source benchmark version 2.
20 models · 20 results · updated 2026-05-04 · vals.ai ↗
Leaderboard
Grok 4.3 xhigh
Heavy line: best score at each price. Thin lines join one model's effort levels. Cost uses each model's median subscription plan, or API price with no plan data. 11 more models under All models in settings.
More from this benchmark
1 published measuresTime per task54 results · lower is better
14 more models under All models in settings.
All results
20 rows| # | Model | Effort | CaseLaw (Vals) | Subscription $ / task | API $ / task | Time |
|---|---|---|---|---|---|---|
| 1 | Grok 4.3 | xhigh | 79.3% | $0.040 | $0.040 | 43s |
| 2 | Opus 4.7 | max | 68.4% | $0.16 | $0.16 | 18s |
| 3 | GPT-5.5 | xhigh | 66.2% | $0.004 | $0.16 | 25s |
| 4 | Sonnet 4.6 | max | 64.0% | $0.14 | $0.14 | 36s |
| 5 | Gemini 2.5 Pro | high | 63.9% | $0.26 | $0.26 | 44s |
| 6 | Muse Spark | 63.1% | $0.007 | $0.007 | 62s | |
| 7 | Kimi K2.6 | max | 61.2% | $0.013 | $0.026 | 46s |
| 8 | MiniMax M2.7 | high | 60.9% | $0.001 | $0.003 | 22s |
| 9 | Qwen 3.5 Flash | max | 55.9% | $0.005 | $0.005 | 44s |
| 10 | MiniMax M2.1 | 55.8% | $0.008 | $0.008 | 37s | |
| 11 | GLM 4.7 | none | 54.9% | $0.021 | $0.021 | 60s |
| 12 | MiniMax M2.5 | high | 53.5% | $0.006 | $0.012 | 31s |
| 13 | Qwen 3.6 27B | max | 53.2% | $0.024 | $0.024 | 42s |
| 14 | Gemma 4 31B IT | high | 52.6% | free | 29s | |
| 15 | GLM 5.1 | max | 51.6% | $0.009 | $0.034 | 49s |
| 16 | Qwen 3.6 Plus | max | 51.4% | $0.023 | $0.047 | 44s |
| 17 | gpt-oss-120b | high | 48.8% | $0.004 | $0.004 | 68s |
| 18 | Qwen 3 Max | max | 47.5% | $0.051 | $0.051 | 85s |
| 19 | Mistral Medium 3.5 | high | 44.2% | $0.004 | $0.085 | 65s |
| 20 | gpt-oss-20b | high | 43.8% | $0.002 | $0.002 | 86s |