Compare
The Kimi K2.7 API, 20 to 25% under list.
Moonshot AI's agentic coding model through the same OpenAI-compatible API as the rest of the catalog, from $0.76 per million input tokens on the Base tier. Two deployments: standard for batch work, fast for interactive sessions.
| Per 1M tokens | Moonshot list | Draftworks Base | Draftworks Scale |
|---|---|---|---|
| Kimi K2.7 Codekimi-k2.7-code | |||
| Input | $0.95 | $0.76 | $0.713 |
| Cached input | $0.19 | $0.152 | $0.142 |
| Output | $4.00 | $3.20 | $3.00 |
| Kimi K2.7 Code Fastkimi-k2.7-code-fast | |||
| Input | $1.05 | $0.84 | $0.787 |
| Cached input | $0.18 | $0.144 | $0.135 |
| Output | $4.40 | $3.52 | $3.30 |
List is Moonshot AI's published price for each deployment class. Base applies from your first dollar; Scale applies permanently at $5,000 lifetime purchases. The same discount covers every model in the catalog.
Common questions.
- How much does the Kimi K2.7 API cost on Draftworks?
- kimi-k2.7-code is $0.76 per 1M input tokens, $0.152 per 1M cached input, and $3.20 per 1M output on the Base tier, against Moonshot AI's list of $0.95, $0.19, and $4.00. The fast deployment is $0.84 input and $3.52 output. The Scale tier pays 25% under list.
- What is the difference between standard and fast mode?
- Both serve the same K2.7 Code weights. The fast deployment runs on a serving stack tuned for interactive latency and costs about 10% more; the standard deployment is the cheaper default for batch and background work.
- Is this the real Kimi K2.7 Code?
- Yes. K2.7 Code is an open-weights model served from Azure, responses report the model id, and every response carries an Ed25519-signed receipt over the model id, token counts, and hashes of the exact bytes in and out.
- Which endpoint does Kimi use?
- POST /v1/chat/completions with the standard OpenAI request shape, including streaming and tool calls. The context window is 256K tokens on both deployments.
- Do I need a Moonshot AI account?
- No. One Draftworks API key covers both Kimi deployments alongside the GPT-5.6 family and Grok 4.3, billed from one prepaid credit balance.