GitHub Copilot token rates
All values below are US dollars per one million tokens. The usage date selects the schedule. GitHub’s pricing table was checked on November 21, 2026. GitHub Copilot model pricing
| Usage date | Uncached input | Cached input | Output |
|---|---|---|---|
| Nov 17, 2026 through May 16, 2027 | $0.05 | $0.005 | $0.25 |
| From May 17, 2027, as scheduled | $0.10 | $0.01 | $0.50 |
One million all-cached input tokens cost half of one US cent during the promotional period. A million output tokens cost 25 cents in that period. For five million uncached input tokens and one million output tokens, the model usage comes to 50 cents before the change and one dollar after it. If a period crosses the change, calculate the usage on each side separately before adding the results.
What does the estimate cover?
This rate card covers MAI Code 2 Flash in GitHub Copilot. Microsoft also announced a limited-access Foundry preview. Your Foundry deployment, account credits, subscription or negotiated terms can have a different billing basis. Use your provider’s billing record when reconciling an invoice. Microsoft launch announcement · GitHub launch details
How do the model limits affect a budget?
The model card gives a maximum of 922,000 input tokens and 128,000 output tokens per call. The upper bound is capacity for one call. Your budget should use the observed tokens for a successful task and account for repeated calls, retries and tests. Microsoft model card
Estimate a specific workload or compare the two MAI Code 2 models.