Kimi K2 Thinking
by Charm Hyper · family kimi-thinking · listed Nov 2025RTW
Thinking Kimi model for slower research passes, planning, and hard technical questions
Current rates USD per 1M tokens
Input$0.60
Output$2.50
Cache read$0.30
Cache write—
Against the market
On the record
- Model id
kimi-k2-thinking- Context window
- 262K tokens
- Max output
- 26K tokens
- Input modalities
- text
- Output modalities
- text
- Reasoning
- yes
- Tool calls
- yes
- Open weights
- yes
- Knowledge cutoff
- Aug 2024
- Released
- 2025-11-06
- Last updated
- 2025-11-06
- Provider docs
- hyper.charm.land
Back of the envelope
≈
Also listed at
- OpenCode$0.40 in · $2.50 out
- Helicone$0.48 in · $2.00 out
- Alibaba (China)$0.574 in · $2.29 out
- 302.AI$0.575 in · $2.30 out
- Qiniu$0.58 in · $2.32 out
- IO.NET$0.60 in · $2.50 out
- Hugging Face$0.60 in · $2.50 out
- Novita AI$0.60 in · $2.50 out
- OpenRouter$0.60 in · $2.50 out
- Kilo Gateway$0.60 in · $2.50 out
- Meganova$0.60 in · $2.60 out
- NanoGPT$0.60 in · $2.50 out
- AI/ML API$0.635 in · $2.65 out
- Poe— in · — out