Filter by provider
Pricing Table
Click column headers to sort. All prices in USD per 1M tokens.
| Model | Input $/1M▲ | Output $/1M | Cached $/1M | Context |
|---|---|---|---|---|
GoogleGemini 1.5 Flash-8BCheapest | $0.037 | $0.150 | $0.0037 | 1M |
GoogleGemini 2.0 Flash-Lite | $0.075 | $0.300 | $0.0075 | 1M |
GoogleGemini 1.5 Flash | $0.075 | $0.300 | $0.0075 | 1M |
GoogleGemini 2.0 Flash | $0.100 | $0.400 | $0.025 | 1M |
GoogleGemini 2.5 Flash-Lite | $0.100 | $0.400 | $0.010 | 1M |
DeepSeekDeepSeek V4 Flash | $0.140 | $0.280 | $0.020 | 1M |
OpenAIGPT-4o Mini | $0.150 | $0.600 | $0.075 | 128K |
GroqLlama 4 Maverick | $0.200 | $0.600 | — | 128K |
AnthropicClaude 3 Haiku | $0.250 | $1.25 | $0.030 | 200K |
GoogleGemini 2.5 Flash | $0.300 | $2.50 | $0.030 | 1M |
DeepSeekDeepSeek V4 Pro | $0.435 | $0.870 | $0.0036 | 1M |
OpenAIGPT-5.4 Mini | $0.750 | $4.50 | $0.075 | 270K |
AnthropicClaude 3.5 Haiku | $0.800 | $4.00 | $0.080 | 200K |
AnthropicClaude Haiku 4.5 | $1.00 | $5.00 | $0.100 | 200K |
OpenAIo4-mini | $1.10 | $4.40 | $0.275 | 200K |
GoogleGemini 1.5 Pro | $1.25 | $5.00 | $0.125 | 2M |
GoogleGemini 3.5 Flash | $1.50 | $9.00 | $0.150 | 1M |
OpenAIo3 | $2.00 | $8.00 | $0.500 | 200K |
OpenAIGPT-5.4 | $2.50 | $15.00 | $0.250 | 270K |
OpenAIGPT-4o | $2.50 | $10.00 | $1.25 | 128K |
AnthropicClaude 3.7 Sonnet | $3.00 | $15.00 | $0.300 | 200K |
AnthropicClaude Sonnet 4.6 | $3.00 | $15.00 | $0.300 | 200K |
OpenAIGPT-5.5 | $5.00 | $30.00 | $0.500 | 270K |
AnthropicClaude Opus 4.8 | $5.00 | $25.00 | $0.500 | 200K |
AnthropicClaude 3 Opus | $15.00 | $75.00 | $1.50 | 200K |
Monthly Cost Projection
Estimated Monthly Cost
Frequently Asked Questions
›How often is the pricing data updated?
We verify pricing from official provider pages regularly. Each model shows its last-verified date. LLM pricing changes frequently — always double-check the provider's official page before making purchase decisions.
›What does "Cached $/1M" mean?
Some providers offer discounted pricing when input tokens are served from a cache. Anthropic, OpenAI, and Google all charge 10% of base input for cached reads. DeepSeek charges just 2%. This can dramatically reduce costs for repeated prompts.
›Why is Groq missing cached pricing?
Groq (Llama 4 Maverick) uses a high-throughput inference architecture that does not currently offer separate cached token pricing. All requests are charged at the standard input rate.
›Which model is cheapest for production use?
For high-volume tasks, DeepSeek V4 Flash ($0.14/1M input) and Gemini 2.5 Flash-Lite ($0.10/1M input) are the most affordable. For quality-sensitive work, Claude 3.5 Haiku ($0.80/1M) or GPT-5.4 Mini ($0.75/1M) offer a good balance.