The pricing is 40% of Kimi K3 at the same Model size. It will be interesting to see benchmarks tomorrow and to see how the pricing evolves across this size class.
We're reaching a point where API cost is no longer the bottleneck for agentic workflows
I think the next 6 month will be a golden age for inference. The weights will be available for anyone to download from hugging face which could create a price fall once providers start competing for your business.
Also: Qwen 3.8 27B coming next week!
Also: Qwen 3.8 Max is competitive with CURRENT best models from OpenAI and Anthropic (and Google, and Facebook, insofar they're even still in the race)