Superintelligence Labs line from Meta reaching frontier tier: AA Intelligence Index 62.1, up from 56.8 on version 1.2 in under a month.
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 59 models released within six months of it, across 3 benchmarks where scores spread out.
| Benchmark | Score | Rank |
|---|---|---|
AA Intelligence (Historical) Archived release-era composite, not comparable to the current versioned Intelligence Index used for rankings | 62.1% | #1 / 22 |
aaAgentic | 59.3% | #1 / 16 |
AA Coding (Historical) Archived release-era coding composite, excluded from current rankings because comparable current-series coverage is unavailable | 76.3% | #7 / 22 |
GPQAArtificial Analysis PhD-level science questions even experts struggle with | 93.5% | #11 / 95 |
HLEArtificial Analysis Challenging multidisciplinary questions evaluated by Artificial Analysis | 48.7% | #12 / 95 |
| Benchmark / source | Score |
|---|---|
HLE Artificial AnalysisMuse Spark 1.3 (Max) | 48.7% |
Terminal-Bench 2.1 Artificial AnalysisMuse Spark 1.3 (Max) | 84.3% |
Terminal-Bench 4.0 Artificial AnalysisMuse Spark 1.3 (Max) | 33.3% |
SciCode Artificial AnalysisMuse Spark 1.3 (Max) | 58.8% |
AA-LCR Artificial AnalysisMuse Spark 1.3 (Max) | 83% |
CritPt Artificial AnalysisMuse Spark 1.3 (Max) | 24.9% |
ITBench SRE Artificial AnalysisMuse Spark 1.3 (Max) | 33.2% |