The Kimi K3 architecture is a scaled-up production version of the Kimi Linear model, increasing from 48 billion parameters to 2.8 trillion parameters. A new component in Kimi K3 is the LatentMoE, which enhances its capabilities as the largest open-weight model currently available.
sebastianraschka.com
2 min
7/28/2026
Kimi-K3 is the latest model from Moonshot AI, set for release on July 27, 2026. It is the world's first open 3T-class model, featuring a new architecture based on Kimi Delta Attention and Attention Residuals, with native agentic capabilities including tool calling, browsing, and multi-step reasoning.
huggingface.co
1 min
7/27/2026
Moonshot AI's Kimi K3 was launched on July 17 but halted new subscriptions due to overwhelming demand. The company plans to release K3's full model weights by July 27, enabling other organizations to host and modify the model.
scientificamerican.com
5 min
7/24/2026
Kimi K3, a 2.8T parameter model from Moonshot AI, scores 57 on the Artificial Analysis Intelligence Index and has an Elo of 1543 on the AA-Briefcase benchmark, making it second only to Fable 5. Kimi K3 costs more to run than Opus 4.8 and averages nearly an hour per task, showing a +727 improvement over Kimi K2.6.
artificialanalysis.ai
3 min
7/22/2026
Kimi K3 is an open model that offers competitive performance with Fable 5 at a lower cost. When combined, Kimi K3 and Fable 5 achieve state-of-the-art results across approximately 1,030 agentic tasks.
fireworks.ai
4 min
7/21/2026
Moonshot Labs launched the Kimi K3 model, while Alibaba introduced the Qwen 3.8 model, both of which are reportedly close in performance to Anthropic's Fable 5. These models will have their weights released publicly in the coming weeks, posing a strategic challenge to leading model developers.
emergingtrajectories.com
5 min
7/20/2026
Kimi K3 delivers performance comparable to Claude in coding tasks, achieving similar output quality and token counts. K3's API pricing is significantly lower at $3 per million input tokens and $15 per million output tokens, while Claude's top model costs more.
stephen.bochinski.dev
3 min
7/18/2026
Moonshot AI has launched Kimi K3, their most capable model to date with 2.8 trillion parameters. It is currently accessible via their website and API, with an open weight release scheduled for July 27, 2026.
simonwillison.net
6 min
7/17/2026
The Kimi K3 architecture is a scaled-up production version of the Kimi Linear model, increasing from 48 billion parameters to 2.8 trillion parameters. A new component in Kimi K3 is the LatentMoE, which enhances its capabilities as the largest open-weight model currently available.
sebastianraschka.com
2 min
7/28/2026
Kimi-K3 is a project hosted on GitHub by MoonshotAI that focuses on collecting and integrating user feedback. The documentation provides details on available qualifiers for user input processing.
github.com
1 min
7/27/2026
Moonshot AI's Kimi K3 was launched on July 17 but halted new subscriptions due to overwhelming demand. The company plans to release K3's full model weights by July 27, enabling other organizations to host and modify the model.
scientificamerican.com
5 min
7/24/2026
Kimi K3 is an open model that offers competitive performance with Fable 5 at a lower cost. When combined, Kimi K3 and Fable 5 achieve state-of-the-art results across approximately 1,030 agentic tasks.
fireworks.ai
4 min
7/21/2026
Kimi K3 delivers performance comparable to Claude in coding tasks, achieving similar output quality and token counts. K3's API pricing is significantly lower at $3 per million input tokens and $15 per million output tokens, while Claude's top model costs more.
stephen.bochinski.dev
3 min
7/18/2026
Kimi K3, the first 2.8T parameter open-source model, is now on Telnyx Inference. 1M context window, native vision, configurable reasoning. Input $2.70/1M tokens.
telnyx.com
1 min
7/27/2026
Kimi-K3 is the latest model from Moonshot AI, set for release on July 27, 2026. It is the world's first open 3T-class model, featuring a new architecture based on Kimi Delta Attention and Attention Residuals, with native agentic capabilities including tool calling, browsing, and multi-step reasoning.
huggingface.co
1 min
7/27/2026
Kimi K3, a 2.8T parameter model from Moonshot AI, scores 57 on the Artificial Analysis Intelligence Index and has an Elo of 1543 on the AA-Briefcase benchmark, making it second only to Fable 5. Kimi K3 costs more to run than Opus 4.8 and averages nearly an hour per task, showing a +727 improvement over Kimi K2.6.
artificialanalysis.ai
3 min
7/22/2026
Moonshot Labs launched the Kimi K3 model, while Alibaba introduced the Qwen 3.8 model, both of which are reportedly close in performance to Anthropic's Fable 5. These models will have their weights released publicly in the coming weeks, posing a strategic challenge to leading model developers.
emergingtrajectories.com
5 min
7/20/2026
Moonshot AI has launched Kimi K3, their most capable model to date with 2.8 trillion parameters. It is currently accessible via their website and API, with an open weight release scheduled for July 27, 2026.
simonwillison.net
6 min
7/17/2026
The Kimi K3 architecture is a scaled-up production version of the Kimi Linear model, increasing from 48 billion parameters to 2.8 trillion parameters. A new component in Kimi K3 is the LatentMoE, which enhances its capabilities as the largest open-weight model currently available.
sebastianraschka.com
2 min
7/28/2026
Kimi-K3 is the latest model from Moonshot AI, set for release on July 27, 2026. It is the world's first open 3T-class model, featuring a new architecture based on Kimi Delta Attention and Attention Residuals, with native agentic capabilities including tool calling, browsing, and multi-step reasoning.
huggingface.co
1 min
7/27/2026
Kimi K3 is an open model that offers competitive performance with Fable 5 at a lower cost. When combined, Kimi K3 and Fable 5 achieve state-of-the-art results across approximately 1,030 agentic tasks.
fireworks.ai
4 min
7/21/2026
Moonshot AI has launched Kimi K3, their most capable model to date with 2.8 trillion parameters. It is currently accessible via their website and API, with an open weight release scheduled for July 27, 2026.
simonwillison.net
6 min
7/17/2026
Kimi K3, the first 2.8T parameter open-source model, is now on Telnyx Inference. 1M context window, native vision, configurable reasoning. Input $2.70/1M tokens.
telnyx.com
1 min
7/27/2026
Moonshot AI's Kimi K3 was launched on July 17 but halted new subscriptions due to overwhelming demand. The company plans to release K3's full model weights by July 27, enabling other organizations to host and modify the model.
scientificamerican.com
5 min
7/24/2026
Moonshot Labs launched the Kimi K3 model, while Alibaba introduced the Qwen 3.8 model, both of which are reportedly close in performance to Anthropic's Fable 5. These models will have their weights released publicly in the coming weeks, posing a strategic challenge to leading model developers.
emergingtrajectories.com
5 min
7/20/2026
Kimi-K3 is a project hosted on GitHub by MoonshotAI that focuses on collecting and integrating user feedback. The documentation provides details on available qualifiers for user input processing.
github.com
1 min
7/27/2026
Kimi K3, a 2.8T parameter model from Moonshot AI, scores 57 on the Artificial Analysis Intelligence Index and has an Elo of 1543 on the AA-Briefcase benchmark, making it second only to Fable 5. Kimi K3 costs more to run than Opus 4.8 and averages nearly an hour per task, showing a +727 improvement over Kimi K2.6.
artificialanalysis.ai
3 min
7/22/2026
Kimi K3 delivers performance comparable to Claude in coding tasks, achieving similar output quality and token counts. K3's API pricing is significantly lower at $3 per million input tokens and $15 per million output tokens, while Claude's top model costs more.
stephen.bochinski.dev
3 min
7/18/2026
No more articles to load