Kimi K3
Benchmarks aside, it is the model we find strongest on complex, multi-step tasks and security reviews. Unlike GLM 5.3, it reads images, so it is the pick when a hard problem comes with screenshots or diagrams.
- 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. Starred scores are reported by the publisher on its model card; the rest are independent measurements. About the scores
- DeepSeek V4.1 Flash45
- GLM 5.345
- Kimi K344
- GLM 5.3 Flash42
- DeepSeek V4 Flash34
- Qwen3.6 35B A3B18
- DeepSeek V4.1 Flash77%* reported by the publisher
- GLM 5.369%
- Kimi K368.5%
- GLM 5.3 Flash63.4%
- DeepSeek V4 Flash54.4%* reported by the publisher
Not published: Qwen3.6 35B A3B.
- DeepSeek V4.1 Flash90.6%* reported by the publisher† measured before the 1017 update
- Kimi K385%
- GLM 5.3 Flash84.3%
- GLM 5.383.9%
- DeepSeek V4 Flash78.7%
- Qwen3.6 35B A3B44.9%
† DeepSeek V4.1 Flash: measured on its previous weights, before the 1017 update. No 2.1 score yet for 1017; on Terminal-Bench 4.0 the update took it from 26.8% to 33.8% (Artificial Analysis).
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-k3Or consider
- DeepSeek V4.1 Flash$0.15 in · $0.60 out per 1MBest for fast agentic coding at a low price.DeepSWE 77%* reported by the publisher
- GLM 5.3$1.40 in · $4.40 out per 1MBest for complex, long-horizon coding.DeepSWE 69%
- GLM 5.3 Flash$0.15 in · $0.50 out per 1MBest for everyday coding at a flash price.DeepSWE 63.4%