Ranking
Models with a verification score of at least 75%, ordered by median output tokens per second. Top 15 of 16.
| # | Model | Speed (tok/s) | Score | First token | Cost / question |
|---|---|---|---|---|---|
| 1 |  Gemini 3.5 Flash LiteGoogle | 409 | 75.0% | 1.1 s | $0.0034 |
| 2 |  Muse Spark OpenMeta | 220 | 77.5% | 0.5 s | $0.0045 |
| 3 | Muse WonderMeta | 175 | 78.8% | 16.1 s | $0.021 |
| 4 | Gemini 3.8 FlashGoogle | 142 | 82.5% | 3.2 s | $0.0037 |
| 5 |  Claude Opus 5.5Anthropic | 129 | 85.0% | 2.3 s | $0.015 |
| 6 | Claude Sonnet 5.5Anthropic | 126 | 78.8% | 0.7 s | $0.0054 |
| 7 | Gemini 3.1 Pro PreviewGoogle | 124 | 77.5% | 3.4 s | $0.017 |
| 8 |  GPT-5.6 TerraOpenAI | 101 | 75.0% | 1.1 s | $0.0063 |
| 9 | GPT-5.3-CodexOpenAI | 95 | 82.5% | 1.4 s | $0.01 |
| 10 |  Grok 4.7SpaceXAI | 77 | 80.0% | 1.9 s | $0.025 |
| 11 | Claude Fable 5.1Anthropic | 70 | 81.2% | 2.1 s | $0.016 |
| 12 | Gemini 4 ArgonGoogle | 69 | 84.8% | 3.3 s | $0.0063 |
| 13 | GPT-6 AstraOpenAI | 49 | 76.2% | 1.3 s | $0.013 |
| 14 |  Qwen3.8 Max (0902)Qwen | 46 | 75.0% | 1.2 s | $0.011 |
| 15 | GPT-6.1 SolOpenAI | 41 | 76.2% | 2.1 s | $0.0029 |