Research
19 posts
Kimi K3
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.
PerceptionBench
PerceptionBench evaluates visual perception at the level of atomic capabilities discovered from model failures, measuring what multimodal models actually see…
Kimi K2.6
Kimi K2.6 advances open-source coding, featuring long-horizon coding, coding-driven design, agent swarms, proactive agents, and the Claw Groups research prev…
Agent Swarm
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 …
WorldVQA
WorldVQA is a benchmark designed to evaluate atomic vision-centric world knowledge in Multimodal Large Language Models (MLLMs).
Kimi K2.5
Kimi K2.5 defines Visual Agentic Intelligence. Trained on 15T tokens, it introduces SOTA visual coding and autonomous agent swarm. Read the full tech blog.
Kimi Vendor Verifier
Rebuilding the "Chain of Trust" -- Kimi Vendor Verifier.
Kimi K2 Thinking
Kimi K2 Thinking, Moonshot's best open-source thinking model.
Kimi-K2-Instruct-0905
[Read on huggingface.co](https://huggingface.co/moonshotai/Kimi-K2-Instruct-0905)
Kimi K2
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…
Kimi-Researcher
[Read on moonshotai.github.io](https://moonshotai.github.io/Kimi-Researcher/)
Kimi-Dev
open-source coding LLM for software engineering tasks
Kimi-Audio
Kimi-Audio, an open-source audio foundation model excelling in audio understanding, generation, and conversation
Kimina-Prover Preview
Technical report of Kimina-Prover Preview.
Kimi-VL
Kimi-VL: Mixture-of-Experts Vision-Language Model for Multimodal Reasoning, Long-Context Understanding, and Strong Agent Capabilities
Muon is Scalable for LLM Training
Muon is Scalable for LLM Training.
MoBA: Mixture of Block Attention for Long-Context LLMs
MoBA: Mixture of Block Attention for Long-Context LLMs
Kimi K1.5: Scaling Reinforcement Learning with LLMs
[Read on github.com](https://github.com/MoonshotAI/Kimi-k1.5)
Mooncake: A KVCache-centric Disaggregated Architecture for LLM Serving
Mooncake is the serving platform for Kimi, a leading LLM service provided by Moonshot AI. - kvcache-ai/Mooncake
