Frontier reasoning model that preceded GPT-5.6, with strong coding, agentic, and ARC-AGI performance.
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 83 models released within six months of it, across 7 benchmarks where scores spread out.
| Benchmark | Score | Rank |
|---|---|---|
Terminal Agentic terminal coding tasks requiring multi-step execution | 85.6% | #3 / 58 |
AA Coding (Historical) Archived release-era coding composite, excluded from current rankings because comparable current-series coverage is unavailable | 76.4% | #6 / 22 |
MMMUvals.ai College-level multimodal reasoning across 30+ disciplines | 88.3% | #8 / 51 |
GPQA PhD-level science questions even experts struggle with | 93.5% | #9 / 96 |
ARC-AGIARC Prize Novel reasoning tasks requiring fluid intelligence | 90.6% | #9 / 40 |
OSWorld Computer use in real desktop environments | 47.5% | #12 / 13 |
AA Intelligence (Historical) Archived release-era composite, not comparable to the current versioned Intelligence Index used for rankings | 54.8% | #16 / 22 |
MMLU-Provals.ai Harder 10-option successor to MMLU; more reasoning-focused | 88.1% | #21 / 61 |
HLEArtificial Analysis Challenging multidisciplinary questions evaluated by Artificial Analysis | 45.8% | #24 / 102 |
LiveCodeBenchvals.ai Contamination-free competitive programming (filtered by cutoff date) | 85.3% | #27 / 62 |
| Benchmark / source | Score |
|---|---|
HLE Artificial AnalysisGPT-5.5 (Xhigh) | 45.8% |
MMMU-Pro Artificial AnalysisGPT-5.5 (Xhigh) | 79.9% |
Terminal-Bench 2.1 Artificial AnalysisGPT-5.5 (Xhigh) | 84.3% |
Terminal-Bench 4.0 Artificial AnalysisGPT-5.5 (Xhigh) | 14.6% |
SciCode Artificial AnalysisGPT-5.5 (Xhigh) | 55.8% |
AA-LCR Artificial AnalysisGPT-5.5 (Xhigh) | 84.3% |
CritPt Artificial AnalysisGPT-5.5 (Xhigh) | 27.1% |
ITBench SRE Artificial AnalysisGPT-5.5 (Xhigh) | 45.8% |
IFBench Artificial AnalysisGPT-5.5 (Xhigh) | 75.9% |