Meta's next Superintelligence Labs iteration, lifting AA Intelligence Index from 50.6 to 56.8.
| Benchmark | Score | Rank |
|---|---|---|
hleArtificial Analysis | 45.5% | #6 / 76 |
aaAgentic | 49.3% | #8 / 12 |
AA Intelligence Independent composite of agentic work, coding, scientific reasoning, and general capability | 56.8% | #9 / 18 |
AA Coding Independent composite of implementation, terminal use, and real-codebase agent performance | 72.2% | #12 / 18 |
MMMUvals.ai College-level multimodal reasoning across 30+ disciplines | 86.1% | #17 / 57 |
MMLU-Provals.ai Harder 10-option successor to MMLU; more reasoning-focused | 88.3% | #18 / 58 |
GPQAArtificial Analysis PhD-level science questions even experts struggle with | 90.4% | #31 / 88 |