1M context with parallel agent teams, #1 on Arena Elo at 1504
SI v3 uses weighted percentages from one Artificial Analysis configuration. Ranking requires all five benchmarks. Partial scores average only available weights. Scores are not comparable to SI v2 or a probability of reaching singularity.
Compared with 52 models released within six months of it, across 9 benchmarks where scores spread out.
| Benchmark | Score | Rank |
|---|---|---|
SWE-bench Real-world GitHub issue resolution | 80.8% | #3 / 39 |
OSWorld Computer use in real desktop environments | 72.7% | #5 / 13 |
Arena Elo Human preference ranking via blind comparisons | 1504 | #9 / 51 |
HumanEval Coding ability - generating correct Python functions | 95% | #10 / 50 |
Terminal Agentic terminal coding tasks requiring multi-step execution | 65.4% | #12 / 58 |
MMLU-Provals.ai Harder 10-option successor to MMLU; more reasoning-focused | 89.1% | #16 / 61 |
MMLU Tests knowledge across 57 subjects from STEM to humanities | 91% | #17 / 54 |
ARC-AGI Novel reasoning tasks requiring fluid intelligence | 68.8% | #17 / 40 |
MATH Competition-level mathematics problems | 93% | #22 / 50 |
MMMUvals.ai College-level multimodal reasoning across 30+ disciplines | 83.9% | #26 / 51 |
LiveCodeBenchvals.ai Contamination-free competitive programming (filtered by cutoff date) | 84.7% | #29 / 62 |
GPQA PhD-level science questions even experts struggle with | 91.3% | #30 / 96 |
HLEArtificial Analysis Challenging multidisciplinary questions evaluated by Artificial Analysis | 39.9% | #42 / 100 |
| Benchmark / source | Score |
|---|---|
HLE Artificial AnalysisClaude Opus 4.6 (Max) | 39.9% |
MMMU-Pro Artificial AnalysisClaude Opus 4.6 (Max) | 75.4% |
AA-LCR Artificial AnalysisClaude Opus 4.6 (Max) | 78% |
CritPt Artificial AnalysisClaude Opus 4.6 (Max) | 12.6% |
IFBench Artificial AnalysisClaude Opus 4.6 (Max) | 53.1% |