MAI Code 2 in GitHub Copilot: price, plans and credits
Updated 2026-11-18. From GitHub's docs and Microsoft's announcement.
MAI Code 2 costs $0.50 per million input tokens and $2.50 per million output tokens in GitHub Copilot until May 17, 2027, then $1 and $5. MAI Code 2 Flash is $0.05 and $0.25. Both are on every paid Copilot plan (Pro, Pro+, Max, Business, Enterprise); Free gets Flash only through Auto.
MAI Code 2 costs $0.50 per million input tokens and $2.50 per million output tokens in GitHub Copilot, a 50% launch price that runs until May 17, 2027, per GitHub's models and pricing page. After that it doubles to $1 and $5. MAI Code 2 Flash is a tenth of that: $0.05 in and $0.25 out now, $0.10 and $0.50 later.
Microsoft released both models on Nov 17, 2026 at Ignite, per its MAI Code 2 announcement. Everything below is confirmed on Microsoft's and GitHub's own pages, read Nov 18, 2026, except the per-request math, which is our estimate and labeled that way.
MAI Code 2 price in GitHub Copilot
Copilot bills every model by the token and converts the total into GitHub AI Credits, where 1 credit is $0.01. Here are the official per-million rates for Microsoft's coding models:
| Model | Input per 1M | Output per 1M |
|---|---|---|
| MAI Code 2 (promo to May 17, 2027) | $0.50 | $2.50 |
| MAI Code 2 (from May 17, 2027) | $1 | $5 |
| MAI Code 2 Flash (promo) | $0.05 | $0.25 |
| MAI Code 2 Flash (from May 17, 2027) | $0.10 | $0.50 |
| MAI-Code-1.1-Flash | $0.20 | $1.20 |
Cached input is $0.025 per million on MAI Code 2 and $0.005 on Flash during the promo, then $0.05 and $0.01. GitHub lists no cache-write charge and no long-context tier for either model. If you're on an older annual Pro or Pro+ plan that still uses request-based billing, GitHub uses model multipliers instead of these rates.
Which Copilot plans get MAI Code 2
Every paid plan. GitHub's supported models table lists both MAI Code 2 and MAI Code 2 Flash as included on Pro, Pro+, Max, Business and Enterprise. Copilot Free and Student only get Auto model selection, and Auto can route to MAI Code 2 Flash but not to MAI Code 2.
| Copilot plan | AI Credits a month | MAI Code 2 | MAI Code 2 Flash |
|---|---|---|---|
| Free | None published | No | Only via Auto |
| Pro, $10 | $15 | Yes | Yes |
| Pro+, $39 | $70 | Yes | Yes |
| Max, $100 | $200 | Yes | Yes |
Plan prices and credit amounts are from GitHub's Copilot plans page, checked 2026-11-18. The credit figure includes the flex amount each plan adds on top of its base.
How many MAI Code 2 requests you get per month
A lot more than with the big models. This is our estimate for one agent request that sends 50,000 input tokens with no cache hits and gets 5,000 tokens back, priced at GitHub's official rates. Real requests vary, and cache hits make every row cheaper. $15 is Pro's monthly credit allowance; GPT-6.2 Sol and Claude Opus 5.5 need Pro+ or higher.
| Model | Credits per request | Requests per $15 (est.) |
|---|---|---|
| MAI Code 2 (promo to May 17, 2027) | 3.75 | about 400 |
| MAI Code 2 (from May 17, 2027) | 7.5 | about 200 |
| MAI Code 2 Flash (promo) | 0.375 | about 4,000 |
| MAI Code 2 Flash (from May 17, 2027) | 0.75 | about 2,000 |
| MAI-Code-1.1-Flash | 1.6 | about 937 |
| GPT-6 Luna | 0.75 | about 2,000 |
| Claude Haiku 5.5 | 0.75 | about 2,000 |
| GPT-6.1 Terra | 7.5 | about 200 |
| GPT-6.2 Sol | 15 | about 100 |
| Claude Sonnet 5.5 | 15 | about 100 |
| Claude Opus 5.5 | 30 | about 50 |
So at the promo price, MAI Code 2 runs about four times as many of these requests as GPT-6.2 Sol or Claude Sonnet 5.5, and eight times as many as Claude Opus 5.5. MAI Code 2 Flash costs about half of GPT-6 Luna or Claude Haiku 5.5 per request. Once the promo ends, MAI Code 2 is still half the price of GPT-6.2 Sol.
Is MAI Code 2 good enough to switch to?
For routine coding, Microsoft says yes, but that's Microsoft's own testing. Its announcement claims MAI Code 2 reaches 98% of GPT-6.1 Sol's score at about 40% of the cost, and 96% of Claude Opus 5.5's at 5%, averaged over four coding benchmarks run in the Copilot CLI at max reasoning effort. Those cost figures use the 50% promo price.
Microsoft also says plainly that both models are "explicitly not yet tuned for harder problems, like fundamental science research and very long horizon agentic work." No independent leaderboard we use ranks them yet, so treat the scores as reported, not confirmed. Our take: make MAI Code 2 your default for everyday edits and fixes while the promo runs, and switch to a frontier model when a task needs long, multi-step agent work.
Can you use MAI Code 2 outside Copilot?
Not yet. MAI Code 2 is only in GitHub Copilot. MAI Code 2 Flash is also in a limited-access preview on Microsoft Foundry, which has its own billing, and Microsoft says Flash is coming "soon" to OpenRouter, Vercel, Hermes, Open Code and Kilo. Neither model has a public API price, so it isn't in our API price table.
Sources
FAQ
How much does MAI Code 2 cost?
$0.50 per million input tokens and $2.50 per million output tokens in GitHub Copilot until May 17, 2027, then $1 and $5, per GitHub's models and pricing page. Copilot converts that into AI Credits at $0.01 each.
Is MAI Code 2 free in GitHub Copilot?
No. It draws from your plan's monthly AI Credits like any other model. Copilot Free can't pick it; Free's Auto mode can route to MAI Code 2 Flash.
Which Copilot plans include MAI Code 2?
Pro, Pro+, Max, Business and Enterprise, per GitHub's supported models table. Free and Student get Auto model selection only.
What is the difference between MAI Code 2 and MAI Code 2 Flash?
MAI Code 2 is the bigger model for multi-file work and debugging. Flash is a tenth of the price, at $0.05 in and $0.25 out during the promo, and Microsoft aims it at explanations, targeted edits and quick fixes.
When does the MAI Code 2 promo price end?
May 17, 2027. From then GitHub lists MAI Code 2 at $1 in and $5 out per million tokens, and Flash at $0.10 and $0.50.
Can I use MAI Code 2 through an API?
Not yet. MAI Code 2 is only in GitHub Copilot. MAI Code 2 Flash is in a limited-access preview on Microsoft Foundry, and Microsoft says OpenRouter and others are coming soon.
Related
- GitHub Copilot pricing and changes
- GitHub Copilot limits explained
- Copilot Free vs Pro vs Pro+ vs Max
- Is Copilot Pro worth it?
- Copilot usage limit reached
- Cheapest AI API for coding
- Which AI models each plan includes
More guides
- Claude Fable 5.2 in Claude Code: limits and cost
- Meta Muse Code limits explained
- Muse Code usage draining fast? Meta's fixes and resets
- Muse Code usage limit reached: what to do
- Claude Code weekly limit: how it works on Pro, Max and Ultra
- Claude Code limits explained
Not sure a plan beats the API for you? Try the plan vs API cost calculator.