# OpenRouter Rankings

Shows which AI models people actually use most on OpenRouter, a service that lets developers switch between many AI models. It reflects real-world popularity, not test scores.

Category: performance. Type: leaderboard. Status: active.
Metric: Tokens (pieces of text, in billions) processed by the model in the latest weekly period (higher is better).
Source: OpenRouter — https://openrouter.ai/rankings, results as of 2026-11-16, retrieved 2026-11-17, license: not stated.

| Rank | Model | Tokens (B) | Requests (M) |
|---|---|---|---|
| 1 | deepseek/deepseek-v4.1-flash-20261017 | 61170.1 | 978.5 |
| 2 | tencent/hy4-20261104 | 16042.2 | 115.5 |
| 3 | deepseek/deepseek-v4.1-flash-20260910 | 15455.0 | 495.9 |
| 4 | z-ai/glm-5.3-flash-20260826 | 11806.2 | 361.0 |
| 5 | openai/gpt-6-luna-20260922 | 8081.3 | 614.8 |
| 6 | anthropic/claude-opus-5.5-20260921 | 7365.4 | 63.3 |
| 7 | stepfun/step-5-preview-20261008 | 6360.0 | 24.8 |
| 8 | xiaomi/mimo-v2.6-flash-20260921 | 4164.6 | 59.1 |
| 9 | z-ai/glm-5.3-20260816 | 3738.6 | 81.6 |
| 10 | openai/gpt-6.1-sol-20260929 | 3711.0 | 78.7 |
| 11 | minimax/minimax-m3.1-flash-20261019 | 3350.9 | 46.9 |
| 12 | moonshotai/kimi-k3.1-pro-20261023 | 3241.9 | 39.8 |
| 13 | nvidia/nemotron-3-ultra-550b-a55b-20260604 (free) | 3198.3 | 37.8 |
| 14 | deepseek/deepseek-v4-flash-20260731 | 3196.4 | 305.6 |
| 15 | nvidia/nemotron-3.5-super-vl-20261015 (free) | 2969.1 | 27.0 |
| 16 | meta/muse-spark-1.3-contributor-20260902 | 2964.9 | 34.3 |
| 17 | typesafe/jev-1.13-20260917 | 2868.8 | 956.9 |
| 18 | anthropic/claude-haiku-5.5-20261007 | 2779.1 | 229.7 |
| 19 | anthropic/claude-sonnet-5.5-20260928 | 2666.3 | 63.9 |
| 20 | baidu/ernie-5.5-flash-preview-20261110 (free) | 2470.8 | 19.1 |

Notes: Weekly view (bucket dated 2026-11-16), data embedded in the rankings page. Tokens = prompt + completion tokens, ranked by that sum (sorted by this sum). Only top 20 models in the page payload. Popularity/usage measure, not quality.
