Every model. Real prices, zero markup.
Compare transparent pay-as-you-go token prices across 9 providers · 100 models, USD per 1M tokens. Live from /api/pricing.
OpenAI
25 modelsAnthropic
10 modelsAlibaba
19 modelsDeepSeek
2 modelsMoonshot
3 modelsMiniMax
1 modelZ.ai
4 modelsByteDance
13 modelsPricing notes
- All prices are in USD. Billing is based on actual token usage.
- Input tokens = your prompt. Output tokens = the model’s response.
- Tiered pricing: a request’s entire token count is charged at the rate of the tier its input falls into. Larger inputs cross over to a higher tier.
- Cache hit: when the same prompt prefix is reused (context caching), input tokens are billed at the discounted Cache/1M rate shown above.
- Image and audio generation models are billed per API call.
- Volume discounts available for enterprise customers. Contact us.
Found your model? Start calling it in 60 seconds.
Get your API key, freeFrequently asked questions
How do I compare LLM API prices across providers?
Synthorai's models page is a live price comparison across every major provider — sort by input or output price per million tokens, filter by provider, always in sync with the gateway's actual list prices. Every figure is exactly what a request draws from your balance, with a 0% platform surcharge.
Which LLM API providers does Synthorai support?
Synthorai is a multimodal AI gateway with one API and one bill for GPT (OpenAI), Claude (Anthropic), Gemini (Google), Qwen (Alibaba), DeepSeek, and more — across chat, image, and audio, with native prompt caching, BYOK, and zero data retention by default. The table above lists every model with its live price.
Which LLM API is the cheapest?
The table above is sorted by list price, so you can pick the cheapest model that meets your quality bar. Prompt caching bills cache hits at about 10% of the input price, and there is no platform fee — so effective cost is often lower than an aggregator at the same list price.