Ranking
Models with a verification score of at least 75%, ordered by billed cost per question (reasoning tokens included). Top 15 of 17.
| # | Model | Cost / question | Cost / correct answer | Score | Speed (tok/s) |
|---|---|---|---|---|---|
| 1 |  GPT-6.1 SolOpenAI | $0.0029 | $0.0037 | 76.2% | 41 |
| 2 |  Gemini 3.5 Flash LiteGoogle | $0.0034 | $0.0045 | 75.0% | 409 |
| 3 | Gemini 3.8 FlashGoogle | $0.0037 | $0.0045 | 82.5% | 142 |
| 4 |  Muse Spark OpenMeta | $0.0045 | $0.0058 | 77.5% | 220 |
| 5 | GPT-6.1 Terra ProOpenAI | $0.005 | $0.0065 | 76.2% | — |
| 6 |  Claude Sonnet 5.5Anthropic | $0.0054 | $0.0069 | 78.8% | 126 |
| 7 | Gemini 4 ArgonGoogle | $0.0063 | $0.0074 | 84.8% | 69 |
| 8 | GPT-5.6 TerraOpenAI | $0.0063 | $0.0084 | 75.0% | 101 |
| 9 | GPT-5.3-CodexOpenAI | $0.01 | $0.013 | 82.5% | 95 |
| 10 |  Kimi K3.1 ProMoonshotAI | $0.01 | $0.014 | 77.5% | 38 |
| 11 |  Qwen3.8 Max (0902)Qwen | $0.011 | $0.015 | 75.0% | 46 |
| 12 | GPT-6 AstraOpenAI | $0.013 | $0.017 | 76.2% | 49 |
| 13 | Claude Opus 5.5Anthropic | $0.015 | $0.017 | 85.0% | 129 |
| 14 | Claude Fable 5.1Anthropic | $0.016 | $0.02 | 81.2% | 70 |
| 15 | Gemini 3.1 Pro PreviewGoogle | $0.017 | $0.022 | 77.5% | 124 |