Many users affirm that OpenRouter data on surging Chinese AI model usage is directionally correct while noting it misses substantial activity inside agents, coding tools, and internal systems.
Based on 7 visible X reactions from 29 accounts; directional sample.
Ask a question below.
Published answers will appear here.
@emollick Agreed, and it's not just Codex/Cowork siphoning share, API usage from internal tools never touches these routers at all. Any metric built on a single aggregator is going to systematically undercount whoever's biggest customers are doing direct integrations.
@joshuaduniganx @DKThomp @pkedrosky And thanks for the ramp data, we're a tiny tiny contributor to those lines:)
@DaveRekuc @DKThomp @pkedrosky Lol nice catch, thanks Dave
@DaveRekuc @DKThomp @pkedrosky 🙏💪
I think OpenRouter is not a good measure of actual model usage in a world of agentic tools (not that I doubt that Chinese open weights model usage is up, but this could also look like a graph of usage shifting to Codex/Code/Cowork). We really need better data indicators for AI! https://twitter.com/DKThomp/status/2076733479274815687
The share peaked at 47%, signalling rising open-weight adoption.
@emollick Agreed, and it's not just Codex/Cowork siphoning share, API usage from internal tools never touches these routers at all. Any metric built on a single aggregator is going to systematically undercount whoever's biggest customers are doing direct integrations.
@joshuaduniganx @DKThomp @pkedrosky And thanks for the ramp data, we're a tiny tiny contributor to those lines:)
I think OpenRouter is not a good measure of actual model usage in a world of agentic tools (not that I doubt that Chinese open weights model usage is up, but this could also look like a graph of usage shifting to Codex/Code/Cowork). We really need better data indicators for AI! https://twitter.com/DKThomp/status/2076733479274815687
OpenRouter is at most a few percentage points of all AI usage. So these charts showing open weight models gaining share on OR don't represent AI market share overall. But they do indicate a directional move towards open weight AI. Canary in a coal mine, perhaps. https://twitter.com/DKThomp/status/2076733479274815687
since February, Chinese AI share of tokens used by US companies has quadrupled to almost 50% (Open Router data, via @pkedrosky) https://x.com/DKThomp/status/2076733479274815687/photo/1
Many users affirm that OpenRouter data on surging Chinese AI model usage is directionally correct while noting it misses substantial activity inside agents, coding tools, and internal systems.
Based on 7 visible X reactions from 29 accounts; directional sample.
Ask a question below.
Published answers will appear here.
OpenRouter is at most a few percentage points of all AI usage. So these charts showing open weight models gaining share on OR don't represent AI market share overall. But they do indicate a directional move towards open weight AI. Canary in a coal mine, perhaps. https://twitter.com/DKThomp/status/2076733479274815687