Every model, its price, and what your plan includes
Each model draws your plan’s monthly pool at its real cost, so the Standard models go furthest. Past the pool, a model costs its provider’s price plus 10%, and your own key costs nothing from us.
Showing 14 of 14 models
| Can do | ||||
|---|---|---|---|---|
| Canopy Standard | ||||
GPT-6 LunaOpenAI | InputCachedOutput$0.11$0.011$0.55 | 1.02.6 uncached | 1.1M | |
DeepSeek V4.1 Flash 1017DeepSeek | InputCachedOutput$0.33$0.007$1.32 | 2.17.6 uncached | 1M | |
| Named only | ||||
Claude Haiku 5.5Anthropic | InputCachedOutput$0.11$0.011$0.55 | 1.02.6 uncached | 1M | |
DeepSeek V4.1 FlashDeepSeek | InputCachedOutput$0.33$0.007$1.32 | 2.17.6 uncached | 1M | |
Gemini 3.5 Flash LiteGoogle | InputCachedOutput$0.33$0.033$2.75 | 4.09.0 uncached | 1M | |
Gemini 3.8 FlashGoogle | InputCachedOutput$0.825$0.083$4.125 | 7.219.8 uncached | 1M | |
Claude Sonnet 5.5Anthropic | InputCachedOutput$2.20$0.11$11.00 | 17.352.8 uncached | 1M | |
Gemini 4 ArgonGoogle | InputCachedOutput$2.20$0.11$11.00 | 17.352.8 uncached | 1M | |
GPT-6.2 SolOpenAI | InputCachedOutput$2.20$0.11$11.00 | 17.352.8 uncached | — | |
Grok 4.7xAI | InputCachedOutput$2.20$0.55$6.60 | 20.448.4 uncached | 500K | |
Kimi K3.1 ProMoonshot AI | InputCachedOutput$2.53$0.253$12.65 | 22.060.7 uncached | 1M | |
Claude Opus 5.5Anthropic | InputCachedOutput$4.40$0.22$22.00 | 34.5105.6 uncached | 1M | |
GPT-6.2 AstraOpenAI | InputCachedOutput$8.80$0.44$44.00 | 69.1211.2 uncached | — | |
GPT-6.1 TerraOpenAI | Price coming | — | 1.1M | |
Prices apply past your plan’s pool. Pool draw says how many Canopy tokens one token costs, so a higher number empties the pool faster. Where a model’s price differs by provider, the highest is shown.
How pricing works
One pool. Every model draws it at its real cost.
- 01
Your plan’s pool
Every plan includes a monthly pool of Canopy tokens. A light model draws it slowly, a frontier model faster, and you only ever watch one number.
- 02
Past the pool
A model costs its provider’s price plus 10%, per 1M tokens, as listed on its card.
- 03
Your own key
On models that take your own provider key, the model costs nothing from us. Open a model’s details to see if it does.
tokens a monthplan poolpool drawGood to know
What is a Canopy token?
The unit your plan’s monthly pool is counted in. 1M Canopy tokens is $0.05 of included usage. Each model draws the pool at its own rate, shown as “Pool draw” on its card.
How is “tokens a month” worked out?
Your plan’s pool divided by the model’s Canopy tokens per token. The counts assume 85% of each prompt is cached. Switch “Show” to a plan to see it for every model.
What does “cached” mean?
Cached input is prompt text the provider already has on hand, so it is billed at a lower rate than fresh input. Long conversations and agent loops are mostly cached.
What are house modes?
A house mode picks a model for you from its own list. Models listed as Named only are the ones you choose yourself.
- Canopy Standard: Everyday work at the lowest cost.
Where are the models served from?
Open Details on any model to see the hosts and regions it runs on, and who developed it.
Ready to pick a plan?
Compare plans side by side, then come back to see exactly how far each pool goes.
See plans and pricing