MODEL PRICES
LLM model prices
Input, output, cached reads, cache writes and audio rates for every model we price, per million tokens. Re-synced every day from each provider's published schedule. This is the same sheet that prices every event in your dashboard.
| PROVIDER | CACHED | CACHE WRITE | REASONING | AUDIO IN | AUDIO OUT | CONTEXT | MODALITIES | |||
|---|---|---|---|---|---|---|---|---|---|---|
| amazon.nova-2-lite-v1:0 | amazon-bedrock | $0.33 | $2.75 | $0.08 | $0.33 | — | — | — | 1M | textimagevideopdf |
| amazon.nova-lite-v1:0 | amazon-bedrock | $0.06 | $0.24 | $0.01 | $0.06 | — | — | — | 300K | textimagevideopdf |
| amazon.nova-micro-v1:0 | amazon-bedrock | $0.04 | $0.14 | $0.00875 | $0.04 | — | — | — | 128K | text |
| amazon.nova-pro-v1:0 | amazon-bedrock | $0.8 | $3.2 | $0.2 | $0.8 | — | — | — | 300K | textimagevideopdf |
| anthropic.claude-fable-5 | amazon-bedrock | $10 | $50 | $1 | $12.5 | — | — | — | 1M | textimagepdf |
| anthropic.claude-fable-5-1 | amazon-bedrock | $10 | $50 | $0.25 | $12.5 | — | — | — | 1M | textimagepdf |
| anthropic.claude-haiku-4-5-20251001-v1:0 | amazon-bedrock | $1 | $5 | $0.1 | $1.25 | — | — | — | 200K | textimagepdf |
| anthropic.claude-haiku-5-5tiered | amazon-bedrock | $0.1 | $0.5 | $0.01 | $0.13 | — | — | — | 1M | textimagepdf |
| anthropic.claude-opus-4-1-20250805-v1:0 deprecated | amazon-bedrock | $15 | $75 | $1.5 | $18.75 | — | — | — | 200K | textimagepdf |
| anthropic.claude-opus-4-5-20251101-v1:0 | amazon-bedrock | $5 | $25 | $0.5 | $6.25 | — | — | — | 200K | textimagepdf |
A dash means the provider publishes no separate rate for that component. We never substitute an invented rate. Where a rate is missing and we must approximate, the affected request in your dashboard is marked as using a derived rate rather than a published one.
How to read these rates
Every figure is per million tokens. A request is priced component by component: text input, cached reads, cache writes, audio in, reasoning, text out and audio out each bill at their own rate.
Cached reads cost a fraction of the input rate, and the discount is not a constant — it varies by model. Cache writes cost more than ordinary input, so caching can raise a bill as well as lower one.
Audio tokens are priced separately from text tokens on the same model and usually cost more. Image generation bills as output tokens, which is why image models carry a high output rate.
These rates price every event in LLMtrack. The free plan uses the same engine. Start tracking free →