openrouter top 5 by token volume this week

a stealth model dominated the openrouter leaderboard. ox alpha – revealed as @Zai_org's glm-5.3 flash – hit 27.2t tokens

1. ox alpha (glm-5.3 flash) – 27.2t
2. deepseek v4 flash – 12.5t
3. xiaomi mimo-v2.5 – 11t
4. tencent hy3 – 6.83t
5. nvidia nemotron 3 ultra (free) – 5.52t

pricing across the chinese top 4: $0.068 to $0.126 per million input tokens. glm-5.3 flash at $0.075/m – currently 50% off.

deepseek holds 3 spots in the top 10. anthropic – zero. gpt-5.6 luna sits at #7, down 13%.

gemini 3.7 flash is the fastest-growing non-chinese model – up 183% to #8. google is quietly clawing back volume while everyone watches the price war.

a model listed under "stealth" routed more tokens than deepseek, openai and google combined. when the price-to-quality ratio is right, builders don't need to know who's behind the api. z-ai just proved you can win the volume game before anyone even knows your name.