Moonshot AI's Kimi K3 opens on Amazon Bedrock with 2.8 trillion parameters
Moonshot AI announced that its Kimi K3 model is now available on Amazon Bedrock. The model is described as the first open‑weight model to reach 2.8 trillion parameters and adds native vision support with a 1-million-token context window. Moonshot AI says Kimi K3 offers an approximate 2.5x improvement in scaling efficiency over its predecessor Kimi K2, making it suitable for long‑running coding…
Key points
- Kimi K3 is the first open‑weight model with 2.8 trillion parameters and a 1-million-token context window
- Explicit prompt caching reduces latency and input‑token costs, with cached content kept for at least 30 minutes
- Global cross‑Region inference costs approximately 10% less than a geographic profile
Amazon Bedrock provides explicit prompt caching for Kimi K3, allowing developers to mark reusable prompt prefixes after at least 1,024 tokens. Cached tokens are billed at a higher rate but remain in cache for at least 30 minutes, and subsequent matching requests receive discounted input‑token pricing and no input‑tokens‑per‑minute quota impact. Global cross‑Region inference with the model costs approximately 10% less than a geographic profile, while a US‑only profile satisfies data‑residency requirements. The model can be accessed via the Bedrock console, the OpenAI‑compatible APIs, or through open‑source agents such as OpenCode and Hermes that support Bedrock providers.
Model page: Kimi K3 →
The story so far
4 episodes →- Moonshot AI's Kimi K3 opens on Amazon Bedrock with 2.8 trillion parametersthis story
Introducing Kimi K3 on Amazon Bedrock
AWS Machine Learning Blog · 18 September 2026
Loading the full article…
This text was published by AWS Machine Learning Blog and written by Alex Thewsey. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
Coverage and discussion
2sources- Moonshot AI’s Kimi K3 Arrives on Amazon Bedrock With 1M-Token ContextPress · Unite.AI ·
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Generative AI & Models
All →- Introducing Mistral Large 4 · 7 src
- OpenAI boosts GPT-6 speed by 50% in 28 Days update · 1 src
- Reflection AI debuts Beam, a 501B open-weight model rivaling GLM 5.2 with 3-4x less compute · 10 src
- Author tests four AI assistants on flawed forecasting scenarios · 1 src
- Anthropic prompts users to share voice data to enhance AI models · 8 src
Comments
via GitHub Discussions