Moonshot AI's Kimi K3.1 Pro Highspeed lands on Amazon Bedrock
Amazon Bedrock now offers Kimi K3.1 Pro Highspeed, a post-trained version of Moonshot AI's Kimi K3 that keeps the same open-weight architecture and 1-million-token context window. AWS says Moonshot used substantially more post-training compute, aimed at long-horizon agentic work and fewer factual errors, along with financial modeling, office documents, and coding for game engines, 3D scenes and CAD scripts. The Highspeed variant emits output tokens faster; Moonshot reports 2–3 times the speed of Kimi K3.1 Pro. It runs on the same endpoint, APIs and features as Kimi K3 through global, US and India inference profiles, and supports tool calling, structured output, configurable reasoning, image input and streaming.
Why it matters
Developers already building with Kimi K3 on Bedrock can try the new model by changing the model ID, and the higher output rate shortens each step of an agentic edit-test-fix loop. AWS also says prompts and completions stay inside its data boundary and are not shared with Moonshot AI.
Sources
- PrimaryAWS Machine Learning
Written by AIToolMaps from the source above and checked against it sentence by sentence.