What each open model costs through every provider that lists it, cheapest first, from public gateway and catalog data. Prices in USD per 1M tokens, checked every hour.
Pareto Inference runs this page and is one of the providers listed. Rows are sorted by price, and Pareto's rows are labeled and placed by price like any other. The methodology says what is included and what is left out. Spot a wrong number? Email us.
Models
GLM 5.3 Flash93 providers. Cheapest now: Pareto Inference (Vercel AI Gateway), $0.01017 on the agent blend.DeepSeek V4.1 Flash72 providers. Cheapest now: Baidu (OpenRouter), $0.005516 on the agent blend.MiMo V2.6 Flash24 providers. Cheapest now: GMICloud (OpenRouter or Vercel AI Gateway), $0.0101 on the agent blend.
How it works
- Sources. OpenRouter's and Vercel AI Gateway's public provider lists, plus models.dev for providers on neither. Every row names its source and when it was checked.
- One unit. USD per 1M tokens, with peak and off-peak prices and promos marked.
- Agent blend first. Rows sort by the price of a coding-agent workload, mostly cached input as measured on Pareto's traffic, with the common 3:1 input-to-output blend beside it.
- Reuse. Each model's table downloads as CSV or JSON.
- Rules. Read the methodology for what is included, how rows are sorted, and how to send a correction.