Short answer
No—not through Moonshot’s official API today. Moonshot’s Kimi API documentation says kimi-k2.5 was retired across its platform on August 31, 2026. That means K2.5 is not a model you can newly budget for or deploy through the official API now, whether paid or free. The documentation points new users to a current model instead. (platform.kimi.ai)
This is a model-availability answer, not proof that no third-party service still offers a K2.5-labeled endpoint. Third-party access has separate availability, billing, and operating terms; confirm those with the provider before relying on them.
What “free Kimi K2.5 API” means
Keep these access routes separate:
- Moonshot’s official API: K2.5 was retired from the platform on August 31, 2026, so there is no current official K2.5 API price or free quota to quote. The official platform provides API access to its available models, but a free app or web-chat experience is not evidence of free API usage. (
platform.kimi.ai) - Third-party API providers: A provider may list or route a K2.5 model, but its endpoint, model version, price, free credits, limits, and production terms are its own offer. Do not treat that as Moonshot’s official API or assume the offer is free.
- Old pricing pages or model listings: These may describe historic rates or a model that is no longer available from Moonshot. They do not establish a price you can buy at now.
The Kimi platform’s pricing documentation describes token billing, including input, output, and cache pricing for models it lists. But because K2.5 is retired, its current official pricing cannot be used to calculate a K2.5 API budget. (platform.kimi.ai)
What should you budget?
For Moonshot’s official K2.5 API, the practical budget is not applicable: the model has been retired, so there is no current official K2.5 rate or quota to calculate against. A reproducible estimate would require a live provider, a confirmed model identifier, that provider’s current input/output rates and any cache rates, plus your expected token volume.
For a current Moonshot API model, use the official pricing page for that exact model and endpoint. A basic estimate is:
estimated cost = (input tokens ÷ 1,000,000 × input rate) + (output tokens ÷ 1,000,000 × output rate)
Then account for any applicable cache or other billable features shown in the provider’s pricing terms. The official platform documents token-based pricing and cache-related pricing, but the applicable rates depend on the model. (platform.kimi.ai)
Example of the calculation, not a K2.5 quote: If a provider lists an input rate of I dollars per million tokens and an output rate of O dollars per million tokens, a workload using 2 million input tokens and 0.5 million output tokens would cost 2 × I + 0.5 × O, before any other billable features. Use the rates from the provider you will actually call; do not insert historical K2.5 rates and present them as current.
Is K2.5 suitable for a production API workload?
Not through Moonshot’s official API as a new deployment: the documented retirement means you should not build a production plan around official K2.5 availability. Choose a model currently listed by the provider and check its own rate limits, billing, and service terms. The official Kimi quickstart describes API access to currently available models, while the retirement notice directs new users to a different model. (platform.kimi.ai)
If a third party still offers K2.5, production suitability depends on that provider’s specific terms—not on the retired Moonshot model listing. Before committing, confirm the exact model ID, continued availability, rate limits, total pricing, and any service commitments directly with that provider. Do not infer production reliability from a model capability description or from the fact that an endpoint accepts API requests.
Bottom line: If you searched for a “Kimi K2.5 free API,” Moonshot’s official K2.5 API is no longer an available option, so there is no current official free quota or price to budget. For production, select a currently available model—or independently validate a third-party K2.5 offer—and calculate costs using that provider’s current terms.
Sources
Use Kimi in OpenClaw — Kimi API Platform