From the source
Together AI has added
Alibaba's flagship Qwen4 Max to its serverless inference, with a 1M-token context window and a 30% discount through November 22, 2026.
- Together says it gave the model a bug report and all 344 files of the Tornado repository in one request, and the patch it returned passes the maintainers' tests.
- The request took 72 seconds, with the first token after 54 seconds, and cost 69 cents at the sale price, or 98 cents at list price.
- Together notes that requests have to stream, and that thinking is on by default, with thinking tokens counted as output tokens.
- Repeating a roughly 630,000-token request cost about 3 cents instead of 66, because Together caches prompt input automatically and nearly all of it came back cached.
- Alibaba released Qwen4 Max on November 16, 2026, and describes it as the most capable model in the Qwen family to date, positioning it for long agent work such as coding across large repositories.






