Moonshot has some of the best coding models. Its K2.6 is up there with some of the best models around. The API for it is now live. It has an input price of $0.16 and $0.95 with cache hit and misses. The output price is $4 for a million tokens. This new model is better for long-term horizon coding. It also is better at instruction following. How does that compare to Claude Sonnet? You are looking at $3 per M tokens (cache miss) and an output price of $15.
📢 Kimi K2.6 API is live
• Input Price (Cache Hit): $0.16 / M tokens
• Input Price (Cache Miss): $0.95 / M tokens
• Output: $4.00 / M tokensKimi K2.6 is our latest + most intelligent model – stronger long-horizon coding, better instruction following & self-correction.… pic.twitter.com/QUDRNAnKbf
— Kimi.ai (@Kimi_Moonshot) April 21, 2026
This model is a multimodal video/text/image AI model with 256k context. It supports tool calls and web search.
[HT]
