AI API prices, per model and per provider
All rates in USD per million tokens, straight from the providers (updated: August 2026). Sort on price, filter on EU providers or open weights, and search your model. These are the upstream prices on which the Gatewaybird API will apply its margin.
| Provider | Class | ||||
|---|---|---|---|---|---|
| GPT-5 | OpenAIUS | Frontier | $1.25 | $10.00 | 400k |
| GPT-5 mini | OpenAIUS | Mid-tier | $0.25 | $2.00 | 400k |
| GPT-5 nano | OpenAIUS | Fast | $0.05 | $0.40 | 400k |
| GPT-4oMultimodaal | OpenAIUS | Frontier | $2.50 | $10.00 | 128k |
| GPT-4o mini | OpenAIUS | Fast | $0.15 | $0.60 | 128k |
| o4-miniReasoning | OpenAIUS | Mid-tier | $1.10 | $4.40 | 200k |
| Claude Opus 4.1 | AnthropicUS | Frontier | $15.00 | $75.00 | 200k |
| Claude Sonnet 4.5 | AnthropicUS | Frontier | $3.00 | $15.00 | 200k |
| Claude Haiku 4.5 | AnthropicUS | Mid-tier | $1.00 | $5.00 | 200k |
| Claude 3.7 Sonnet | AnthropicUS | Mid-tier | $3.00 | $15.00 | 200k |
| Gemini 2.5 Pro | GoogleUS | Frontier | $1.25 | $10.00 | 1000k |
| Gemini 2.5 Flash | GoogleUS | Fast | $0.30 | $2.50 | 1000k |
| Gemini 2.5 Flash-Lite | GoogleUS | Fast | $0.10 | $0.40 | 1000k |
| Llama 4 MaverickOpen gewichten | MetaUS | Open weights | $0.27 | $0.85 | 1000k |
| Llama 4 ScoutOpen gewichten | MetaUS | Open weights | $0.11 | $0.34 | 3000k |
| Mistral Large 2EU-provider | Mistral AIEU | Frontier | $2.00 | $6.00 | 128k |
| Mistral Medium 3EU-provider | Mistral AIEU | Mid-tier | $0.40 | $2.00 | 128k |
| Mistral Small 3.1EU-provider, open gewichten | Mistral AIEU | Fast | $0.10 | $0.30 | 128k |
| CodestralCode, EU-provider | Mistral AIEU | Mid-tier | $0.30 | $0.90 | 256k |
| DeepSeek V3.1 | DeepSeekCN | Open weights | $0.27 | $1.10 | 128k |
| DeepSeek R1Reasoning | DeepSeekCN | Open weights | $0.55 | $2.20 | 128k |
| Grok 4 | xAIUS | Frontier | $3.00 | $15.00 | 256k |
| Grok 4 Fast | xAIUS | Fast | $0.20 | $0.50 | 2000k |
| Qwen3 235BOpen gewichten | AlibabaCN | Open weights | $0.19 | $0.69 | 128k |
| Qwen3 Coder | AlibabaCN | Open weights | $0.29 | $1.19 | 256k |
| Pharia-1 LLM 7BDuitsland | Aleph AlphaEU | Open weights | $0.35 | $0.35 | 32k |
| Tülu 3 70BOpen gewichten | AI2 (hosting EU)EU | Open weights | $0.60 | $0.90 | 128k |
| Gemma 3 27BOpen gewichten | GoogleUS | Open weights | $0.07 | $0.20 | 128k |
28 28 of 28 models · Prices in USD per 1M tokens · Updated: August 2026
Source: official provider rates. Gatewaybird rates are announced at launch.
How to read these prices
Input vs output
Input tokens are what you send, output tokens what the model returns. Output costs 3-5x more at nearly all providers. Budget your use case on both: a chat assistant with long answers weighs heavier on output than a classification task.
What Gatewaybird adds on top
The gateway passes through the provider price plus a margin (indicative 8-15%). The exact margin is announced at launch — for the waitlist the first period is without margin or discounted.
EU, US and China
The provider column shows the country of the provider entity. For GDPR-sensitive use cases the processing location of the model matters: EU providers (Mistral, EU hosting) process inside the EEA, US and CN providers do not.
Open weights
Models with open weights (Llama, Qwen, Gemma, Mistral Small, DeepSeek) can be self-hosted or served by a provider of your choice. The shown prices are indications of common hosted rates; self-hosting can be cheaper.
Be there at launch?
Waitlist members get API access first and influence on which providers get connected.
You are on the waitlist
We will email you as soon as the API goes live.