Benchmark
Orak AceAttorney
Agentic benchmarks measure how well models carry out tasks that take several steps. Scores are percentages. Higher is better.
- Models measured
- 6
- Scores recorded
- 10
- Sources
- 1
- IQ weight
- 0.52
Score distribution
- 66.7%
- Mean
- 81.9%
- Median
- 20.2%
- Std dev
- 49.9%
- 10th percentile
- 85.3%
- 90th percentile
- 28.4%
- Min
This benchmark has a weight of 0.52 for chat models, measured over 6 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 |
|---|---|---|
| 1 | o3 Minidefault· Orak leaderboard docs | 91.7% |
| 2 | GPT 4odefault· Orak leaderboard docs | 85.3% |
| 3 | GPT 4oreflection· Orak leaderboard docs | 85.3% |
| 4 | DeepSeek R1default· Orak leaderboard docs | 83.3% |
| 5 | Claude Sonnet 3.7default· Orak leaderboard docs | 81.9% |
| 6 | Gemini 2.5 Pro Preview 03-25default· Orak leaderboard docs | 55.7% |
| 7 | GPT 4oref-plan· Orak leaderboard docs | 52.8% |
| 8 | GPT 4oplanning· Orak leaderboard docs | 52.7% |
| 9 | GPT 4ozeroshot· Orak leaderboard docs | 49.9% |
| 10 | GPT-4o minidefault· Orak leaderboard docs | 28.4% |