First mainline model incorporating GPT-5.3-Codex coding capabilities. Native computer use, 1M context, surpasses human baseline on OSWorld. Codex branch ends here.
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 62 models released within six months of it, across 10 benchmarks where scores spread out.
| Benchmark | Score | Rank |
|---|---|---|
FrontierMath Unpublished research-level mathematics problems | 47.6% | #1 / 4 |
OSWorld Computer use in real desktop environments | 75% | #3 / 13 |
SWE-bench Real-world GitHub issue resolution | 78.2% | #9 / 39 |
Terminal Agentic terminal coding tasks requiring multi-step execution | 75.1% | #9 / 58 |
MATH Competition-level mathematics problems | 96.7% | #12 / 50 |
HumanEval Coding ability - generating correct Python functions | 94.5% | #13 / 50 |
ARC-AGI Novel reasoning tasks requiring fluid intelligence | 73.3% | #14 / 40 |
GPQA PhD-level science questions even experts struggle with | 92.8% | #17 / 96 |
LiveCodeBenchvals.ai Contamination-free competitive programming (filtered by cutoff date) | 84% | #32 / 62 |
MMMUvals.ai College-level multimodal reasoning across 30+ disciplines | 79.2% | #38 / 51 |
MMLU-Provals.ai Harder 10-option successor to MMLU; more reasoning-focused | 84.6% | #45 / 61 |
HLEArtificial Analysis Challenging multidisciplinary questions evaluated by Artificial Analysis | 28.1% | #70 / 100 |
| Benchmark / source | Score |
|---|---|
HLE Artificial AnalysisGPT-5.4 mini (Xhigh) | 28.1% |
MMMU-Pro Artificial AnalysisGPT-5.4 mini (Xhigh) | 73.3% |
Terminal-Bench 2.1 Artificial AnalysisGPT-5.4 mini (Xhigh) | 59.2% |
Terminal-Bench 4.0 Artificial AnalysisGPT-5.4 mini (Xhigh) | 2% |
SciCode Artificial AnalysisGPT-5.4 mini (Xhigh) | 52.1% |
AA-LCR Artificial AnalysisGPT-5.4 mini (Xhigh) | 77% |
CritPt Artificial AnalysisGPT-5.4 mini (Xhigh) | 10% |
ITBench SRE Artificial AnalysisGPT-5.4 mini (Xhigh) | 35.2% |
IFBench Artificial AnalysisGPT-5.4 mini (Xhigh) | 73.3% |