| # | Model | HLE | Reasoning | Cap. | Cost |
|---|
| 1 |  Muse Spark 1.1 Meta Superintelligence Labs | 52.2% | — | — | $2.00/M |
| 2 |  Claude Fable 5.1 Anthropic | 46.5% | — | — | $20.00/M |
| 3 | | 46.4% | 94.4 | 86.4 | $4.50/M |
| 4 | | 43.5% | 93.5 | — | $3.21/M |
| 5 | | 38.1% | 90.9 | 84.7 | $2.21/M |
| 6 | | 37.5% | 92.6 | 84.8 | — |
| 7 | | 36.2% | 93.3 | 87.4 | $5.63/M |
| 8 | | 34.4% | 88.4 | 86.8 | $10.00/M |
| 9 |  Inkling Thinking Machines Lab | 29.7% | 87.2 | — | $1.76/M |
| 10 | | 25.3% | 86.2 | 81.8 | $3.44/M |
| 11 | | 24.4% | — | 80.3 | $1.00/M |
| 12 | | 20.3% | 81.8 | 79.2 | $3.50/M |
| 13 | | 18.1% | 79.6 | 78.6 | $1.93/M |
| 14 |  Gemini 2.5 Pro (05-06) Google | 17.8% | 66.7 | 75.3 | — |
| 15 | Gemini 3.1 Flash Lite Google | 8.6% | 81.8 | 76.7 | $0.56/M |
| 16 | | 8.3% | — | — | $1.00/M |
| 17 | | 8.1% | — | — | $0.31/M |
| 18 | | 5.4% | 66.9 | 70.4 | $3.50/M |
| 19 | Claude 3.5 Sonnet Anthropic | 4.1% | 55.3 | 67.5 | — |
| 20 | | 2.7% | 49.2 | 63.0 | $4.38/M |
20 of 20 models, ranked by HLE
