Moonshot AI · Kimi Code
Kimi K2.7 Code
Kimi’s dedicated coding model for long-context, tool-using software work, with a separate high-speed endpoint for the same model.
Keep it as a specialist comparison. Run the same repository task against K3 and the current coding tool before accepting the provider’s improvement claim.
Use it for
- Long-context repository work
- Multi-step coding agents
- Code tasks that use image or video evidence
Do not choose a specialist coding model for ordinary knowledge work, or assume the high-speed endpoint improves the answer rather than only its output speed.
Effort without the jargon
The coding task needs multi-step reasoning; this model does not offer a non-thinking mode.
Advanced recordCost, limits, privacy and operations
K2.7 Code is a specialist coding endpoint with a long context window; use it only where the coding workflow can preserve its reasoning state and handle fixed parameters.
- API cost
- Depends on surfaceKimi publishes model-specific token rates in its live pricing pages; the cited quickstart does not give a stable amount to reproduce here.
- Context window
- Officially disclosed256,000 tokens
- Maximum output
- Not disclosedNot disclosed in the cited record
- Knowledge cutoff
- Not disclosedNot disclosed in the cited record
- Recorded inputs
- Text · Images · Video through documented formats
- Recorded output
- Text · Tool calls
Data handling is a product decision
Not disclosedThe cited model and pricing records do not state a portable retention or training rule. Confirm the Kimi API terms before repository or customer data is sent.
Operational constraints
- Non-thinking mode is not supported.
- Tool choice is limited to auto or none, and reasoning content must be retained during multi-step tool calls.
- The high-speed endpoint is the same model with limited capacity and potentially fluctuating service.
Unknown means the cited official record does not disclose a safe value. It is not an estimate. Prices are provider list prices in USD where stated and can change before this record's review date.
Keep the evidence separate
Moonshot positions K2.7 Code as its specialist coding model with 256K context, multimodal tool use and improved long-horizon instruction following over K2.6.
No sufficiently useful independent note has been added yet. That absence is not filled with our guess.
No Human Bit result is claimed; use this as evaluation guidance only.
Not scheduled
No Human Bit result is claimed for this model yet.