benchmarks
Top Models by Benchmark Score (2026)
Ranked by blended benchmark data from Aider Polyglot and Artificial Analysis Intelligence Index. Available models only - any under access restrictions are excluded.
| # | Model | Blended |
|---|---|---|
| 1 | Claude Opus 5 | 100.0 |
| 2 | Anthropic: Claude Fable 5 | 98.4 |
| 3 | OpenAI: GPT-5.6 Sol | 98.2 |
| 4 | MoonshotAI: Kimi K3 | 94.8 |
| 5 | Anthropic: Claude Opus 4.8 | 94.3 |