CAP
Knowledge & QA
League table — best at knowledge & qa · 30
mean percentile across this capability's benchmarks| # | Model | League | Coverage | $/1M | tok/s |
|---|---|---|---|---|---|
| 1 | Gemini 3 Pro Preview | 100.0 | 1/2 | $4.5 | — |
| 2 | Claude Opus 4.5 | 99.9 | 1/2 | $10 | — |
| 3 | Gemini 3 Flash Preview | 99.9 | 1/2 | $1.13 | — |
| 4 | Claude 4.1 Opus | 99.8 | 1/2 | $30 | — |
| 5 | MiniMax-M2.1 | 99.8 | 1/2 | $0.525 | — |
| 6 | Claude 4.5 Sonnet | 99.7 | 1/2 | $6 | — |
| 7 | GPT-5.2 | 99.7 | 1/2 | $4.81 | — |
| 8 | Claude 4 Opus | 99.6 | 1/2 | $30 | — |
| 9 | GPT-5 | 99.5 | 1/2 | $3.44 | — |
| 10 | GPT-5.1 | 99.5 | 1/2 | $3.44 | — |
| 11 | Grok 4 | 99.4 | 1/2 | $6 | — |
| 12 | GPT-5 Codex | 99.4 | 1/2 | $3.44 | — |
| 13 | DeepSeek V3.2 Speciale | 99.3 | 1/2 | — | — |
| 14 | DeepSeek V3.2 | 99.3 | 1/2 | $0.315 | — |
| 15 | Gemini 2.5 Pro | 99.2 | 1/2 | $3.44 | 135 |
| 16 | GPT-5.1 Codex | 99.1 | 1/2 | $3.44 | — |
| 17 | Gemini 2.5 Pro Preview (Mar' 25) | 99.1 | 1/2 | — | — |
| 18 | o1 | 99.1 | 2/2 | $26.25 | — |
| 19 | GLM-4.7 | 99.0 | 1/2 | $1 | — |
| 20 | Grok 4.1 Fast | 99.0 | 1/2 | — | — |
| 21 | Doubao Seed Code | 98.9 | 1/2 | — | — |
| 22 | o3 | 98.9 | 1/2 | $3.5 | 149 |
| 23 | DeepSeek V3.1 Terminus | 98.8 | 1/2 | $1.91 | — |
| 24 | Grok 4 Fast | 98.7 | 1/2 | $0.275 | — |
| 25 | DeepSeek V3.2 Exp | 98.7 | 1/2 | $0.315 | — |
| 26 | DeepSeek R1 0528 (May '25) | 98.6 | 1/2 | $2.06 | — |
| 27 | Cogito v2.1 | 98.6 | 1/2 | $1.25 | — |
| 28 | Kimi K2 Thinking | 98.5 | 1/2 | $1.07 | — |
| 29 | DeepSeek R1 (Jan '25) | 98.5 | 1/2 | $2.43 | — |
| 30 | Qwen3 235B A22B 2507 | 98.4 | 1/2 | $2.63 | — |
League = a model's mean percentile across the 2 benchmarksin this capability (best score per benchmark; partial coverage shown). Price & speed via Artificial Analysis.