Kimi AI

PerceptionBench evaluates visual perception at the level of atomic capabilities discovered from model failures, measuring what multimodal models actually see rather than what they infer.

Kimi K3 is the world's first open 3T-class model — frontier performance across coding, knowledge work, and reasoning, with native multimodality and 1M context.

Kimi K2.6 advances open-source coding, featuring long-horizon coding, coding-driven design, agent swarms, proactive agents, and the Claw Groups research preview.

Kimi Agent Swarm is a self-organizing multi-agent system that spawns up to 100 sub-agents to parallelize research, document synthesis, and multi-perspective analysis.

WorldVQA is a benchmark designed to evaluate atomic vision-centric world knowledge in Multimodal Large Language Models (MLLMs).

Kimi K2.5 defines Visual Agentic Intelligence. Trained on 15T tokens, it introduces SOTA visual coding and autonomous agent swarm.

Rebuilding the "Chain of Trust" -- Kimi Vendor Verifier.

Kimi K2 Thinking, Moonshot's best open-source thinking model.

Kimi K2-Instruct-0905 is the latest, most capable version of Kimi K2. It is a state-of-the-art mixture-of-experts (MoE) language model, featuring 32 billion activated parameters and a total of 1 trillion parameters.

Kimi K2 is our latest Mixture-of-Experts model with 32 billion activated parameters and 1 trillion total parameters. It achieves state-of-the-art performance in frontier knowledge, math, and coding among non-thinking models.

Meet Kimi-Researcher , an autonomous agent that excels at multi-turn search and reasoning. It performs an average of 23 reasoning steps and explores over 200 URLs per task.

We introduce Kimi-Dev-72B, our new open-source coding LLM for software engineering tasks. Kimi-Dev-72B achieves a new state-of-the-art on SWE-bench Verified among open-source models.

We present Kimi-Audio, an open-source audio foundation model excelling in audio understanding, generation, and conversation. This repository contains the official implementation, models, and evaluation toolkit for Kimi-Audio.

Technical report of Kimina-Prover Preview.

We present Kimi-VL, an efficient open-source Mixture-of-Experts (MoE) vision-language model (VLM) that offers advanced multimodal reasoning, long-context understanding, and strong agent capabilities—all while activating only 2.8B parameters in its language decoder (Kimi-VL-A3B).

Recently, the Muon optimizer based on matrix orthogonalization has demonstrated strong results in training small-scale language models, but the scalability to larger models has not been proven. We identify two crucial techniques for scaling up Muon: (1) adding weight decay and (2) carefully adjusting the per-parameter update scale.

Trainable Block Sparse Attention: The full context is divided into blocks, where each query token learns to attend to the most relevant KV blocks, enabling efficient processing of long sequences. Parameter-less Gating Mechanism: A novel Parameter-less top-k gating mechanism is introduced to selects the most relevant blocks for each query token, ensuring that the model focuses only on the most informative blocks.

There are a few key ingredients about the design and training of k1.5.

Mooncake is the serving platform for Kimi, a leading LLM service provided by Moonshot AI.