AI model usage.

S2

stealth/space-bunny-alpha

16.9T

stealth · -49%

1

DeepSeek V4.1 Flash

36.4T

deepseek · +59%

3

GLM 5.3 Flash

10.9T

z-ai · +16%

4

Nemotron 3 Ultra (free)

6.37T

nvidia · tokens+7%

5

GPT-6 Luna (batch)

6.01T

openai · tokens+19%

6

DeepSeek V4 Flash 0731

5.07T

deepseek · tokens-19%

7

Claude Opus 5.5 (batch)

3.79T

anthropic · tokens+117%

8

GLM 5.3 (batch)

3.17T

z-ai · tokens+35%

9

DeepSeek V4 Flash 0423

3.02T

deepseek · tokens-5%

10

MiMo-V2.6-Flash

2.34T

xiaomi · tokens+100%

11

Kimi K3 (batch)

1.87T

moonshotai · tokens+35%

12

Gemini 3.8 Flash (batch)

1.72T

google · tokens-3%

13

Muse Spark 1.3 Contributor

1.63T

meta · tokens+14%

14

GLM 5.2

1.6T

z-ai · tokens+27%

15

GPT-6.1 Sol (batch)

1.57T

openai · tokens+341%

16

GPT-5.6 Luna (batch)

1.51T

openai · tokens-19%

17

GPT-5.6 Sol (batch)

1.5T

openai · tokens+22%

18

Claude Sonnet 5.5 (batch)

1.41T

anthropic · tokens+226%

19

MiMo-V2.6-Pro

1.11T

xiaomi · tokens+5%

20

DeepSeek V4 Pro 0813

1.07T

deepseek · tokens+25%

Source: OpenRouter rankings, 2026-10-03 – 2026-10-09. Licensed under CC BY 4.0. openrouter.ai

How this is ranked · Sources & limitations →