Benchmark
Protocols · Understanding v2
Science benchmarks measure how well models answer science questions. Scores are percentages. Higher is better.
- Models measured
- 6
- Scores recorded
- 9
- Sources
- 2
- IQ weight
- 1.15
Score distribution
- 64.7%
- Mean
- 64.1%
- Median
- 4.1%
- Std dev
- 58.9%
- 10th percentile
- 69%
- 90th percentile
- 58.9%
- Min
This benchmark has a weight of 1.15 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 | Claude Opus 5Opus 5.5 System Card · adaptive thinking · max effort unless stated· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 71.8% |
| 2 | Claude Opus 5.5Opus 5.5 System Card · adaptive thinking · max effort unless stated· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 69% |
| 3 | Claude Sonnet 5.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 66.6% |
| 4 | Claude Sonnet 5.5adaptive thinking · max effort · default sampling · 5-trial mean · context ≤1M unless noted· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 66.6% |
| 5 | Claude Haiku 5.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 64.1% |
| 6 | Claude Haiku 5.5adaptive thinking · max effort · default sampling · 5-trial mean · context ≤1M unless noted· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 64.1% |
| 7 | GPT-6 AstraOpus 5.5 System Card · adaptive thinking · max effort unless stated· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 61.9% |
| 8 | Claude Sonnet 5adaptive thinking · max effort · default sampling · 5-trial mean · context ≤1M unless noted· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 58.9% |
| 9 | Claude Sonnet 5Opus 5.5 System Card · adaptive thinking · max effort unless stated· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it. | 58.9% |
Data credits: BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links. · Data by LLM Stats