Large Language Models
docAnalyzer keeps pace with releases from Alibaba, Anthropic, DeepSeek, Google, Meta, Minimax, Moonshot AI, OpenAI, StepFun, Tencent, Xiaomi, and xAI, adding new models within days so you can pick the right fit for every workflow. We surface 28 models across 12 creators, each benchmarked for quality, speed, and responsiveness.
Refreshed 2026-10-03
57.6 top quality · Claude Opus 5.5
344 tok/s top speed · Gemini 3.5 Flash Lite
850 ms top responsiveness · DeepSeek V4.1 Flash
| Model | |||
|---|---|---|---|
 Claude Opus 5.5 Anthropic | 57.6 | 98 tok/s | 488.33 s |
| Claude Sonnet 5.5 Anthropic | 56.0 | 141 tok/s | 261.12 s |
| Claude Fable 5.1 Anthropic | 53.4 | 71 tok/s | 137.87 s |
 GPT-6 Astra OpenAI | 52.7 | 57 tok/s | 197.03 s |
| GPT-6.1 Sol OpenAI | 51.8 | 59 tok/s | 155.39 s |
 Muse Spark 1.3 Meta | 48.1 | 164 tok/s | 45.91 s |
 Grok 4.7 xAI | 46.4 | 82 tok/s | 22.45 s |
 MiMo V2.6 Pro Xiaomi | 46.3 | 46 tok/s | 3.64 s |
 Qwen3.8 Max Alibaba | 45.4 | 38 tok/s | 1.87 s |
 Kimi K3 Moonshot AI | 43.6 | 36 tok/s | 2.62 s |
 Gemini 3.8 Flash Google | 40.9 | 268 tok/s | 12.49 s |
 DeepSeek V4.1 Flash DeepSeek | 39.5 | 221 tok/s | 850 ms |
| GPT-6 Luna OpenAI | 38.1 | 138 tok/s | 84.21 s |
| MiMo V2.6 Flash Xiaomi | 37.9 | 53 tok/s | 4.24 s |
| GPT-5.6 Luna OpenAI | 37.3 | Not available | Not available |
| Qwen 3.8 27B Alibaba | 33.7 | 47 tok/s | 1.18 s |
| Gemini 3.1 Pro Preview Google | 29.7 | 126 tok/s | 26.72 s |
| Qwen3.7 Max Alibaba | 29.5 | Not available | Not available |
 MiniMax M3 Minimax | 29.2 | 74 tok/s | 1.11 s |
| MiMo V2.5 Pro Xiaomi | 26.0 | Not available | Not available |
Hy3 Tencent | 25.3 | 89 tok/s | 2.08 s |
| Qwen3.7 Plus Alibaba | 25.2 | 56 tok/s | 1.69 s |
| MiMo V2.5 Xiaomi | 25.2 | Not available | Not available |
| Grok 4.3 xAI | 24.9 | Not available | Not available |
| Gemini 3.5 Flash Lite Google | 22.2 | 344 tok/s | 7.17 s |
Step 3.7 Flash StepFun | 19.5 | Not available | Not available |
| Claude Haiku 4.5 Anthropic | 16.9 | 115 tok/s | 14.97 s |
| GPT-5 nano OpenAI | 13.0 | Not available | Not available |
* Latency uses a logarithmic scale: shorter time means a longer bar.