Lower-cost GPT-6 model for coding and everyday work. Independent measurements track max effort; missing evaluations are not estimated.
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 models released within six months of it.
| Benchmark | Score | Rank |
|---|---|---|
ARC-AGIARC Prize Novel reasoning tasks requiring fluid intelligence | 61.8% | #22 / 39 |
HLEArtificial Analysis Challenging multidisciplinary questions evaluated by Artificial Analysis | 38.5% | #44 / 95 |
| Benchmark / source | Score |
|---|---|
HLE Artificial AnalysisGPT-6 Luna (Max) | 38.5% |
MMMU-Pro Artificial AnalysisGPT-6 Luna (Max) | 79.7% |
Terminal-Bench 4.0 Artificial AnalysisGPT-6 Luna (Max) | 12.6% |
SciCode Artificial AnalysisGPT-6 Luna (Max) | 54.6% |
AA-LCR Artificial AnalysisGPT-6 Luna (Max) | 83.3% |
CritPt Artificial AnalysisGPT-6 Luna (Max) | 19.4% |
DeepSWE 1.1 Provider reportOpenAI research/API evaluation; max effort | 66.6% |