3AM MarketerPrompt caching prices by provider
Prompt caching bills the repeated start of a prompt at a lower cached input rate instead of the full input rate. This table puts the two side by side for every model whose provider publishes a cached price, with the cache write rate where the provider charges one. The median discount is 90% off fresh input.
Cached vs fresh input
Tap a column to sort| Anthropic Claude Fable 5.1 | $10 | $0.25 | $12.50 | 98% |
|---|---|---|---|---|
| Anthropic Claude Mythos 5.1 | $10 | $0.25 | $12.50 | 98% |
| DeepSeek deepseek-flash | $0.30 | $0.006 | – | 98% |
| Meta muse-spark-1.2-contributor | $0.10 | $0.002 | – | 98% |
| Meta muse-spark-1.3-contributor | $0.10 | $0.002 | – | 98% |
| DeepSeek deepseek-v4-pro | $1.32 | $0.044 | – | 97% |
| Anthropic Claude Opus 5.5 | $4 | $0.20 | $5 | 95% |
| Anthropic Claude Sonnet 5.5 | $2 | $0.10 | $2.50 | 95% |
| Google Gemini Gemini 4 Argon | $2 | $0.10 | $0.10 | 95% |
| Meta muse-wonder-1.0 | $4 | $0.20 | – | 95% |
| OpenAI gpt-6.1-sol | $2 | $0.10 | $2.50 | 95% |
| OpenAI gpt-6.1-terra | $1 | $0.05 | $1.25 | 95% |
| Anthropic Claude Fable 5 | $10 | $1 | $12.50 | 90% |
| Anthropic Claude Haiku 3.5 | $0.80 | $0.08 | $1 | 90% |
| Anthropic Claude Haiku 4.5 | $1 | $0.10 | $1.25 | 90% |
| Anthropic Claude Haiku 5.5 | $0.10 | $0.01 | $0.125 | 90% |
| Anthropic Claude Mythos 5 | $10 | $1 | $12.50 | 90% |
| Anthropic Claude Opus 4 | $15 | $1.50 | $18.75 | 90% |
| Anthropic Claude Opus 4.1 | $15 | $1.50 | $18.75 | 90% |
| Anthropic Claude Opus 4.5 | $5 | $0.50 | $6.25 | 90% |
| Anthropic Claude Opus 4.6 | $5 | $0.50 | $6.25 | 90% |
| Anthropic Claude Opus 4.7 | $5 | $0.50 | $6.25 | 90% |
| Anthropic Claude Opus 4.8 | $5 | $0.50 | $6.25 | 90% |
| Anthropic Claude Opus 5 | $5 | $0.50 | $6.25 | 90% |
| Anthropic Claude Sonnet 4 | $3 | $0.30 | $3.75 | 90% |
| Anthropic Claude Sonnet 4.5 | $3 | $0.30 | $3.75 | 90% |
| Anthropic Claude Sonnet 4.6 | $3 | $0.30 | $3.75 | 90% |
| Anthropic Claude Sonnet 5 | $2 | $0.20 | $2.50 | 90% |
| Google Gemini Gemini 2.5 Flash | $0.30 | $0.03 | $0.03 | 90% |
| Google Gemini Gemini 2.5 Flash-Lite | $0.10 | $0.01 | $0.01 | 90% |
| Google Gemini Gemini 2.5 Pro | $1.25 | $0.125 | $0.125 | 90% |
| Google Gemini Gemini 3 Flash Preview | $0.50 | $0.05 | $0.05 | 90% |
| Google Gemini Gemini 3.1 Flash-Lite | $0.25 | $0.025 | $0.025 | 90% |
| Google Gemini Gemini 3.1 Pro Preview | $2 | $0.20 | $0.20 | 90% |
| Google Gemini Gemini 3.5 Flash-Lite | $0.30 | $0.03 | $0.03 | 90% |
| Google Gemini Gemini 3.6 Flash | $0.75 | $0.075 | $0.075 | 90% |
| Google Gemini Gemini 3.8 Flash | $0.75 | $0.075 | $0.075 | 90% |
| Mistral Mistral Large 4 | $1.36 | $0.14 | – | 90% |
| Moonshot kimi-k3 | $3 | $0.30 | $3 | 90% |
| OpenAI chat-latest | $5 | $0.50 | – | 90% |
| OpenAI gpt-5.3-codex | $1.75 | $0.175 | – | 90% |
| OpenAI gpt-5.6-cyber | $12.50 | $1.25 | $15.625 | 90% |
| OpenAI gpt-5.6-sol | $4 | $0.40 | $5 | 90% |
| OpenAI gpt-6-astra | $10 | $1 | $12.50 | 90% |
| OpenAI gpt-6-luna | $0.10 | $0.01 | $0.125 | 90% |
| OpenAI gpt-rosalind-discovery | $5 | $0.50 | – | 90% |
| OpenAI gpt-rosalind-research | $5 | $0.50 | – | 90% |
| Meta muse-spark-1.1 | $1.25 | $0.15 | – | 88% |
| Meta muse-spark-1.2 | $1.25 | $0.15 | – | 88% |
| Meta muse-spark-1.3 | $1.25 | $0.15 | – | 88% |
| xAI grok-4.5 | $2 | $0.30 | – | 85% |
| xAI grok-4.20-0309-non-reasoning | $1.25 | $0.20 | – | 84% |
| xAI grok-4.20-0309-reasoning | $1.25 | $0.20 | – | 84% |
| xAI grok-4.20-multi-agent-0309 | $1.25 | $0.20 | – | 84% |
| xAI grok-4.3 | $1.25 | $0.20 | – | 84% |
| Moonshot kimi-k2.6 | $0.95 | $0.16 | – | 83% |
| Moonshot kimi-k2.7-code | $0.95 | $0.19 | – | 80% |
| Moonshot kimi-k2.7-code-highspeed | $1.90 | $0.38 | – | 80% |
| xAI grok-build-0.1 | $1 | $0.20 | – | 80% |
| xAI grok-4.6 | $2 | $0.50 | – | 75% |
| xAI grok-4.7 | $2 | $0.50 | – | 75% |
US dollars per million tokens, standard tier. Discount is cached input against fresh input. Cache write is the default (5 minute) cache, charged when a prompt is first stored; – means the provider does not print a write rate for that model. Each cached price links to the page it was read from.
Sources
Prices read from each provider's page: Anthropic, DeepSeek, Meta, Google Gemini, OpenAI, Mistral, Moonshot, xAI.
Price your own usage in the LLM API cost calculator. More on how we source figures: how we collect our data.