Together AI and Moonshot AI have formed a strategic partnership making Together AI the launch platform for Moonshot's open-weight model releases. Starting with Kimi K3 — a 2.8 trillion parameter sparse Mixture-of-Experts model with native vision support and a 1M token context window — developers get day-zero access via Together AI's US-hosted infrastructure. Kimi K3 introduces two new architectural components: Kimi Delta Attention (KDA) for faster long-context decoding, and Attention Residuals (AttnRes) for improved training efficiency. The partnership offers serverless, provisioned throughput, and dedicated inference options, plus post-training capabilities including LoRA and full-weight fine-tuning. Future Moonshot model releases will also ship on Together AI at launch through the same API.
