Moonshot API pricing

Moonshot AI's Kimi lineup — K2.5, K2.6, and the Code-specialized K2.7 Code — sits entirely in the Budget tier and has built a strong reputation specifically for coding-adjacent tasks at a low price point. See the live table above for current rates.

The Code variant is a real specialization, not just a name

Unlike providers where a "Code" suffix is mostly branding, Kimi K2.7 Code is specifically tuned for code generation and coding-assistant workloads. If your use case is code-heavy, it's worth checking this variant against the general-purpose K2.6 directly rather than assuming the newer numbered release is automatically better for your task — a specialized variant frequently outperforms a newer general model on the specific task it was tuned for.

K2.5 vs. K2.6 vs. K2.7 Code

The numbered progression (K2.5 → K2.6) represents Moonshot's general-purpose line improving over time, while K2.7 Code branches specifically toward coding. Practical guidance: for general chat and summarization, default to the latest general-purpose release (currently K2.6); for coding-specific workloads, check K2.7 Code against it directly on your actual task rather than assuming either wins by default.

Why Moonshot shows up on value-focused shortlists

Kimi models frequently appear near the top of coding-focused cost-vs-quality comparisons — check the current standing on SWE-bench Verified and Aider Polyglot before assuming a more expensive provider is necessary for your coding workload. The combination of Budget-tier pricing with competitive coding benchmark performance is exactly the profile the value-frontier concept is built to surface — Moonshot is a common example of a model that sits on that frontier rather than being dominated by higher-priced alternatives.

What to verify

As with any lower-cost, less-established provider relative to the largest US labs, it's worth explicitly checking rate limits and throughput for your expected volume, and confirming current terms around data handling for your specific use case, rather than assuming parity with larger providers by default.

When Moonshot vs. a competitor

For coding-specific workloads at the Budget tier, compare Kimi K2.7 Code directly against DeepSeek and Qwen's coder-branded variants on the SWE-bench and Aider cost-vs-quality charts. For general-purpose Budget-tier chat, all three are worth comparing side by side on compare before committing.

See current Kimi pricing live above, or check where it ranks on cheapest models.