Cheapest frontier models
Flagship GPT-5, Claude Opus and Sonnet, Gemini Pro, Grok, and o-series rows from this catalog, ranked by input rate. Example request: 1000 input tokens and 500 output tokens.
All pricing ranks · Cheapest LLM API · Cheapest frontier · Cheapest long replies · Long context · Price one prompt · Cost calculator
Ranked by published rates
| # | Model | Provider | Context | Input / 1M | Output / 1M | Example | 10k requests | |
|---|---|---|---|---|---|---|---|---|
| 1 | Gemini 1.5 Pro | 2M | $1.25/1M | $5.00/1M | $0.00375 | $37.5 | Cost | |
| 2 | Gemini 2.0 Pro Exp | 2M | $1.25/1M | $5.00/1M | $0.00375 | $37.5 | Cost | |
| 3 | GPT-5 | OpenAI | 400K | $1.25/1M | $10.00/1M | $0.00625 | $62.5 | Cost |
| 4 | Gemini 2.5 Pro | 1M | $1.25/1M | $10.00/1M | $0.00625 | $62.5 | Cost | |
| 5 | GPT-4.1 | OpenAI | 1M | $2.00/1M | $8.00/1M | $0.006 | $60.00 | Cost |
| 6 | o3 | OpenAI | 200K | $2.00/1M | $8.00/1M | $0.006 | $60.00 | Cost |
| 7 | Claude Sonnet 5 | Anthropic | 1M | $2.00/1M | $10.00/1M | $0.007 | $70.00 | Cost |
| 8 | GPT-5.6 Terra | OpenAI | 1.1M | $2.00/1M | $12.00/1M | $0.008 | $80.00 | Cost |
| 9 | Gemini 3.1 Pro | 1M | $2.00/1M | $12.00/1M | $0.008 | $80.00 | Cost | |
| 10 | Claude Sonnet 4.6 | Anthropic | 1M | $3.00/1M | $15.00/1M | $0.0105 | $105.00 | Cost |
| 11 | Grok 3 | xAI | 131K | $3.00/1M | $15.00/1M | $0.0105 | $105.00 | Cost |
| 12 | Grok 4 | xAI | 256K | $3.00/1M | $15.00/1M | $0.0105 | $105.00 | Cost |
| 13 | Claude Opus 4.6 | Anthropic | 1M | $5.00/1M | $25.00/1M | $0.0175 | $175.00 | Cost |
| 14 | Claude Opus 4.8 | Anthropic | 1M | $5.00/1M | $25.00/1M | $0.0175 | $175.00 | Cost |
| 15 | GPT-5.6 Sol | OpenAI | 1.1M | $5.00/1M | $30.00/1M | $0.02 | $200.00 | Cost |
| 16 | o1 | OpenAI | 200K | $15.00/1M | $60.00/1M | $0.045 | $450.00 | Cost |
Related tools
LLM cost calculator · Cheapest LLM API · Compare model prices · Cache savings · Batch pricing
Related guides
How LLM API pricing works · How prompt cost is calculated · Prompt caching explained · Exact vs Approx
Sources and references
Official documentation used for definitions, counting methods, or rate cards. Always confirm critical budgets on the provider page.
Pricing rank FAQ
- Which models count as frontier here?
- A short curated set: GPT-5, GPT-4.1, o3, o1, Claude Opus 4.x, Claude Sonnet 5 and 4.6, Gemini Pro rows, Grok 4 and Grok 3. Mini, nano, flash, and host copies are on the full cheapest list.
- Why is a cheaper Flash or Mini missing?
- Those sit on Cheapest LLM API. This page is only the flagship class, so you can compare like with like.
- Does cheaper mean better?
- No. Use the calculator and a real prompt before you pick a model for production.