Kimi K3
A solid pick for self-hosting and full data control.
Overview
Kimi K3, from Moonshot AI, is the first open 3T-class model: 2.8 trillion parameters (104B activated), a 896-expert MoE architecture, and a 1-million-token context with native text/image/video understanding. It targets long-horizon agentic work — coding, research, office automation. Real limit: its raw knowledge score (HLE-Full) trails Claude Fable 5. Weights are open under the Kimi K3 licence and can run on European GPU servers, but its scale and Chinese origin call for careful governance.
Skill profile
Not disclosed
Strengths
- Very long context
- Open-source and self-hostable
Limitations
- No native GDPR guarantee
- API pricing not disclosed
Who is it for
- you want to control cost or self-host
- you handle sensitive EU data
Ideal use cases
- Long-horizon autonomous coding
- Deep research automation
- Office task automation
- Video and image analysis
- Multi-tool agent orchestration
Access & availability
Key specifications
Privacy
Run it locally
| Quantization | Disk | RAM / VRAM | Typical hardware |
|---|---|---|---|
| Q4 · recommended | 1596 GB | 1758 GB | Server GPU infra (H100/A100…) |
| Q8 · balanced | 2996 GB | 3298 GB | Server GPU infra (H100/A100…) |
| FP16 · max quality | 5600 GB | 6162 GB | Server GPU infra (H100/A100…) |
Estimates for a moderate context. Long contexts need more RAM (KV cache).
Deploy
Copy-ready commands generated from this card. Adjust context length and GPU count to your hardware.
OpenAI-compatible server for production on NVIDIA GPUs.
pip install vllm
vllm serve moonshotai/Kimi-K3 \
--max-model-len 32768 \
--tensor-parallel-size 2 \
--dtype autoAdvanced data · for expertsArchitecture, modalities, detailed cost, full benchmarks▾
| Arena Elo | — |
| MMLU | — |
| GPQA | — |
| HumanEval | — |
| SWE-Bench | — |
| MATH | — |