From the source
- AWS made
Moonshot AI's Kimi K3.1 Pro Highspeed available on
Amazon Bedrock on November 10, 2026, with a 1-million-token context window through global, US and India inference profiles.
- Moonshot AI says the Highspeed variant generates output two to three times faster than K3.1 Pro, which matters because a single agentic coding task can require many sequential model calls, and a higher output-token rate shortens the generation portion of each step.
- Existing Kimi K3 implementations on Bedrock can switch by changing the model ID, and the new model supports tool calling, structured output, configurable reasoning, image input and streaming through the same endpoint and APIs.
- AWS says inference data stays within the AWS data boundary, is not shared with Moonshot AI and is not used to train the model, with zero data retention always enabled for inference requests.
- The new model carries higher input and output rates than the preceding Kimi K3, though its global cache-read rate is lower than K3's $0.30 per million tokens, and Flex offers inference at half the Standard rate while Priority carries a 75% premium.
- Converse can fail with an internal server error when reasoning content from earlier turns is included in a multi-turn request, and AWS recommends removing prior reasoning blocks for such requests.