Consensus score
56.1
95% interval 41.4 to 69.7
56.1
95% interval 41.4 to 69.7
42th
IQR spread 57.5 pct
7
Across 4 domains
2
Newest observation 107d ago
Median within-board percentile by capability domain across boards with sufficient peer overlap.
Best verified run per board, with confidence intervals and Pareto frontier status.
| Benchmark | Rank | Score | Cost | Status | Observed |
|---|---|---|---|---|---|
| FrontierCode Epoch AI Benchmarking Hub · high · devin-cli | #41 / 51 | 18.8% | Unknown | Standard | Not stated |
| GPQA diamond Epoch AI Benchmarking Hub · max | #42 / 338 | 91%[87.6% to 94.5%] | Unknown | Standard | 2 Aug 2026 |
| ARC-AGI-2 Epoch AI Benchmarking Hub · max | #96 / 299 | 61.4% | Unknown | Standard | Not stated |
| ARC-AGI-2 Epoch AI Benchmarking Hub · high | #115 / 299 | 56% | Unknown | Standard | Not stated |
| Bug Hunt Bench Bug Hunt Bench · Claude Code | #115 / 145 | 14 | $1.52 | Standard | 1 Aug 2026 |
| ARC-AGI-2 Epoch AI Benchmarking Hub · low | #131 / 299 | 46% | Unknown | Standard | Not stated |
| ARC-AGI-2 Epoch AI Benchmarking Hub · none | #235 / 299 | 2.1% | Unknown | Standard | Not stated |
Same benchmark and harness measured across multiple reasoning effort settings.
Epoch AI Benchmarking Hub
2.1
46.0
56.0
61.4
Direct pairwise win rates for DeepSeek V4-Flash 0731 across shared benchmark boards.
4 shared boards
25% win
Head-to-head4 shared boards
0% win
Head-to-head4 shared boards
25% win
Head-to-head4 shared boards
25% win
Head-to-head4 shared boards
0% win
Head-to-head4 shared boards
0% win
Head-to-head4 shared boards
0% win
Head-to-head4 shared boards
50% win
Head-to-head4 shared boards
0% win
Head-to-head4 shared boards
0% win
Head-to-head