LLM API Pricing Comparison
Compare pricing across OpenAI, Anthropic (Claude), Google (Gemini), Mistral, Cohere, and Azure OpenAI.
All prices in USD per 1 million tokens | Cached input pricing shown where available | Data current as of November 2025
Quick Cost Calculator
Loading models…
| Model | Provider | Input Cost | Output Cost | Cached Input | Context Window |
|---|---|---|---|---|---|
| Loading data… | |||||
Key Insights & Special Features
- Cached Input Pricing: OpenAI, Google, and Azure offer reduced rates for cached input tokens (typically 10% of standard input price). This is beneficial for applications that reuse the same context.
- Context Windows: Google’s Gemini models offer the largest context windows (1M tokens), while most others cap at 200K-256K tokens. Larger windows suit long-document processing and complex reasoning tasks.
- Best Value for Flagship Models: OpenAI’s GPT-5 ($1.25 input / $10 output) and Google’s Gemini 2.5 Pro (identical pricing) are competitively priced at the high end. Claude Sonnet 4.5 at $3/$15 is premium but focuses on coding excellence.
- Most Cost-Effective Options: Cohere’s Command R7B ($0.0375 input / $0.15 output) and Mistral’s Small ($0.014 input / $0.042 output) offer the lowest per-token costs for budget-conscious users.
- Batch API Discounts: Azure and Google offer 50% batch API discounts (check their documentation for specific terms). This benefits non-urgent, high-volume workloads.
- Extended Thinking Tokens: Models with reasoning capabilities (like o-series and Gemini with thinking mode) may charge for extended thinking tokens separately—verify with provider documentation.
Comparison by Use Case
- Coding & Complex Tasks: Claude Sonnet 4.5, GPT-5, or Gemini 2.5 Pro. All are competitive but choose based on coding benchmark performance.
- Cost-Sensitive Production: Cohere Command R ($0.15 input / $0.60 output) or Mistral Medium offer a good balance between capability and cost.
- Large Context Processing: Google Gemini Flash (1M tokens, $0.30 input / $2.50 output) is hard to beat for long documents.
- Budget Projects: Mistral Small or Cohere Command R7B for the absolute lowest costs with acceptable quality.
- Enterprise/Compliance: Azure OpenAI offers regional deployment options and data residency guarantees; Google and Azure support batch processing for cost reduction.
Manage Consent
Functional Always active
The technical storage or access is strictly necessary for the legitimate purpose of enabling the use of a specific service explicitly requested by the subscriber or user, or for the sole purpose of carrying out the transmission of a communication over an electronic communications network.
Preferences
The technical storage or access is necessary for the legitimate purpose of storing preferences that are not requested by the subscriber or user.
Statistics
The technical storage or access that is used exclusively for statistical purposes. The technical storage or access that is used exclusively for anonymous statistical purposes. Without a subpoena, voluntary compliance on the part of your Internet Service Provider, or additional records from a third party, information stored or retrieved for this purpose alone cannot usually be used to identify you.
Marketing
The technical storage or access is required to create user profiles to send advertising, or to track the user on a website or across several websites for similar marketing purposes.
