MoonshotAI API Pricing and Measured Speed
We track 9 models from MoonshotAI. Official API prices are shown first where available; OpenRouter is a separate channel. Prices were last checked . Methodology · Source
Subscription limits · Terms tracker · Model prices · What's new
Official API prices
USD per 1M tokens. Direct provider rates are listed first; OpenRouter prices follow as a separate channel.
| Model | Input | Output | Source | Checked |
|---|---|---|---|---|
| MoonshotAI Kimi K3 | $3.00 | $15.00 | Official price list |
Our measured latency and output speed appear on the linked model pages, by route and country, after successful requests on at least two dates.
Prices through OpenRouter
| Model | Input per 1M tokens | Output per 1M tokens | Context (tokens) | Listed since |
|---|---|---|---|---|
| MoonshotAI: Kimi K2 0711 | $0.57 | $2.30 | 131,072 | 11 Jul 2025 |
| MoonshotAI: Kimi K2 0905 | $0.60 | $2.50 | 262,144 | 4 Sep 2025 |
| MoonshotAI: Kimi K2 Thinking | $0.60 | $2.50 | 262,144 | 6 Nov 2025 |
| MoonshotAI: Kimi K2.5 | $0.53 | $3.33 | 262,144 | 27 Jan 2026 |
| MoonshotAI: Kimi K2.6 | $0.65 | $3.41 | 262,144 | 20 Apr 2026 |
| MoonshotAI: Kimi K2.7 Code | $0.60 | $3.50 | 262,144 | 12 Jun 2026 |
| MoonshotAI: Kimi K3.1 Pro | $2.00 | $10.00 | 1,048,576 | 23 Oct 2026 |
| MoonshotAI: Kimi K3 4 of 6 common tasks fit on price and context · quality not tested | $0.19 | $16.00 | 1,048,576 | 16 Jul 2026 |
Batch variants (1)
OpenRouter lists these as separate batch variants at a different price; batch requests are not answered right away, so these prices are kept apart from live prices.
| Model | Input per 1M tokens | Output per 1M tokens | Context (tokens) | Listed since |
|---|---|---|---|---|
| MoonshotAI: Kimi K3 (batch) | $0.92 | $6.00 | 1,048,576 | 16 Jul 2026 |
Unknown means no value was recorded at our latest check; it is never shown as zero. Prices in USD as OpenRouter lists them; we do not add taxes.
How these numbers are produced
Each row is one model at one provider in our ai-models section. Prices are the input and output prices per one million tokens from the latest check of that model; the context length is the value read on the same pass. The table is sorted by our latest output price, lowest first; rows without an output price come last, and after them, greyed, rows of providers whose terms restrict independent comparisons (list of these providers).
A provider gets its own page when we track at least 5 of its models.
Each model links to its page with dates, the source link and the full history.