Public · unit
Model hours.Money and tokens, always together.
The main unit is the work time the backing buys. Amount and tokens always show together, on a smaller line.
Model hour calculator
Enter an amount or a number of model hours, and pick a tier. The three units always show together, with the platform fee on the effective amount.
Models by tier
The proxy only runs these models for a tier. This list is the catalog it enforces, so a model that is not here is not served.
Essential
Models in this class · 5
- @cf/google/gemma-4-26b-a4b-it
- @cf/meta/llama-4-scout-17b-16e-instruct
- @cf/openai/gpt-oss-120b
- @cf/qwen/qwen3-30b-a3b-fp8
- @cf/zai-org/glm-5.3-flash
Advanced
Models in this class · 4
- @cf/deepseek-ai/deepseek-v4-pro-0813
- @cf/moonshotai/kimi-k2.6
- @cf/moonshotai/kimi-k2.7-code
- @cf/zai-org/glm-5.3
Frontier
The three sizes
Showcase card≈ 31,325 hmodel hours$2,000 · ≈ 3.1B tokens
Backing panel
≈ 720 hmodel hours$46 · ≈ 72M tokens
Receipt≈ 390 hmodel hours$25 · ≈ 39M tokensRate at backing time: about 100,000 tokens/h, a published estimate declared Oct 7, 2026 (sample so far: 0 of 50 active hours, 0 of 5 keys)
Rounding
Up to 10 hto the half hour≈ 7.5 h
Up to 100 hto the hour≈ 46 h
Above 100 hto 5 hours≈ 590 h
Where the number comes from
Active hourAn hour in which a key of the tier made at least one request.
Tier rateMedian tokens per active hour over the last 30 days.
Small sampleThe published default applies, marked as an estimate.
At backing time, the rate freezes
Backers see the conversion at the rate of that moment. It is stored with the backing and never changes later.
How to say it
≈ 120 model hourswork capacityestimate