Moonshot AI

Moonshot's next-generation flagship: a 2.8T-parameter Stable LatentMoE (16 of 896 experts active) built on Kimi Delta Attention (from Kimi Linear) and…

Moonshot's next-generation terminal AI coding agent and the headline harness for the Kimi K2.7-Code launch — all of the model's coding/agentic benchma…

Coding-focused agentic model built on Kimi K2.6 — same 1T-total / 32B-active MoE architecture (384 experts, 8 selected + 1 shared, 61 layers, MLA atte…

Moonshot AI's latest frontier model, building on the Kimi K2 family's 1T-total / 32B-active MoE architecture with 262K token context. Optimized for lo…

Proposes PrfaaS (Prefill-as-a-Service), an LLM serving architecture that selectively offloads long-context prefill to standalone, compute-dense prefil…

Authored by the Kimi Team at Moonshot AI (including Su Jianlin), this paper introduces a fundamental rethink of the "residual connection" by replacing…

Open-source multimodal agentic model designed to advance general agentic intelligence. Features joint text-vision optimization and Agent Swarm, a self…

LLM abstraction layer for building agent applications.

Hybrid linear attention architecture (KDA + MLA). 3B active / 48B total MoE model with 75% KV-cache reduction and 6x throughput at 1M context.

Middleware tool for updating model weights dynamically in inference engines.

Landmark 1-trillion parameter MoE model family (32B active) released as open-weight. 384 experts with 8 activated per token. Pre-trained on 15.5T toke…

Benchmarks developed to evaluate an agent's ability to browse the web and synthesize information from multiple sources.

Multi-language SDK (Python, Go, Node.js) for programmatically interacting with Kimi's agentic features.

Command-line agent designed for autonomous coding and system tasks (Python, PyPI kimi-cli). Being wound down in favor of its successor Kimi Code CLI,…

Open-source SWE LLM achieving 60.4% on SWE-bench Verified. Uses agentless training as skill prior for SWE-Agents.

Specialized iteration for autonomous research.

Specialized library written in Go for high-performance AI infrastructure.

Universal audio foundation model for understanding, generation, and conversation. Trained on 13M+ hours of audio data.

Formal reasoning model for Lean 4 theorem proving trained via RL from Qwen2.5-72B. Achieves 80.7% on miniF2F with pass@8192.

Cleaned and corrected version of the popular formal reasoning benchmark, released alongside Kimina-Prover.

Open-source vision-language model with Moonlight MoE LLM (2.8B active / 16B total) and 400M MoonViT encoder. Kimi-VL-Thinking achieves 64.0 on MMMU.

Optimizer based on matrix orthogonalization. Research demonstrating scalability for large language model training.

Lightweight "Pareto-frontier" models trained using the Muon optimizer. Available in 3B and 16B variants.

Mixture of Block Attention mechanism for efficient long-context processing.

Reasoning-focused model using reinforcement learning, claimed to match OpenAI's o1-preview in math and coding. Technical report details scaling RL wit…

A KVCache-centric disaggregated architecture for LLM serving. Won Best Paper at FAST 2025.

Update to the proprietary model supporting a 2-million-character context window.

First model to support 128,000 to 200,000 tokens of lossless context, a world-first at the time. Included moonshot-v1-8k, 32k, and 128k API variants p…