Providers / Novita AI
Novita AI API Pricing
Novita AI is a gateway: it hosts open-weight models such as DeepSeek R1, Qwen, Llama and OpenAI’s gpt-oss and sells API access at its own per-token rates. These are reseller prices, not official prices from the model makers, and we label them that way everywhere on this site.
Every figure below was read from Novita’s public pricing page and is stamped with its verification date. Where a model maker publishes no official API price we could verify, we show none rather than estimate one. Models are listed by price only. Links to Novita are affiliate links (we may earn a commission at no extra cost to you) — see our disclosure.
Novita AI models and prices
USD per 1M tokens, gateway (reseller) prices, sorted by price.
| Model | Input | Output | Cached | Batch | Verified |
|---|---|---|---|---|---|
| OpenAI GPT OSS 20B | $0.04 | $0.15 | — | — | 2026-10-27 |
| OpenAI GPT OSS 120B | $0.05 | $0.25 | — | — | 2026-10-25 |
| Llama 4 Maverick Instruct | $0.27 | $0.85 | — | — | 2026-10-25 |
| DeepSeek R1 0528 | $0.7 | $2.5 | $0.35 | — | 2026-10-25 |
| Qwen3.8 Max | $2 | $6 | $0.25 | — | 2026-10-25 |
Frequently asked questions
What is the cheapest Novita AI API model?
OpenAI GPT OSS 20B, at $0.04 per 1M input tokens and $0.15 per 1M output tokens (verified 2026-10-27). Its cost calculator shows worked examples at realistic volumes.
Are these Novita AI's official prices?
No — these are the gateway’s own resale prices for open-weight models, verified on its public pricing page and dated. They are not the model makers’ official prices.
How do I estimate my Novita AI bill?
Open any model above and use its calculator: enter your input, output and cached token counts plus your call volume, and it computes the per-call and monthly cost from the same verified prices. The homepage calculator also compares models across providers side by side.