Kimi K3
Retiring November 9, 2026Benchmarks aside, it is the model we find strongest on complex, multi-step tasks and security reviews, and unlike GLM 5.3 it reads images. It retires on November 9, 2026: move to Kimi K3.1 Pro, post-trained from K3 with the same 1M context and image input, at a lower price.
- Price per 1M tokens
- $3.00 in$15.00 out
- $0.30 cache read
- Context
- 1M
- 128K max output
- Parameters
- 2.8T
- 104B active
- Vision
- Tool calling
- Reasoning optional, default max
How it ranks
Against every model we serve, best first. About the scores
- Kimi K3.1 Pro46
- DeepSeek V4.1 Flash33
- GLM 5.332
Not ranked, reported on another version: Kimi K3 (Intelligence v4.3: 44).
- DeepSeek V4.1 Flash77%
- Kimi K3.1 Pro75.5%
- GLM 5.369%
- Kimi K368.5%
- Kimi K3.1 Pro67.2%
- DeepSeek V4.1 Flash42.3%
- GLM 5.341.9%
Not ranked, reported on another version: Kimi K3 (Terminal-Bench 2.1: 85%).
Pricing
- Input
- $3.00 per 1M tokens
- Output
- $15.00 per 1M tokens
- Cache read
- $0.30 per 1M tokens
Specs
- Context window
- 1M tokens
- Max output
- 128K tokens
- Parameters
- 2.8T total, 104B active per token
- Vision
- Yes
- Tool calling
- Yes
- Reasoning
- Adjustable (none, low, high, max)
- Default effort
- max
- In production since
- July 31, 2026
Quickstart
Pick your tool; the command already carries this model's id. More tools (Pi, Copilot, Zed, omp) in the setup guides.
OpenCode
Built inLaunch OpenCode on this model with the umans CLI. Umans AI is also built into OpenCode: run /connect and pick it.
curl -fsSL https://api.code.umans.ai/cli/install.sh | bash # once
umans opencode --model umans-kimi-k3
