Microsoft, Anthropic · The makers’ own facts

MAI Code 2 vs Claude Opus 5.5 (2026): Microsoft’s price and benchmark claims, and where each is available

MAI Code 2, Microsoft’s coding model released on November 17, 2026 in GitHub Copilot, costs $0.50 / $2.50 per million input / output tokens at its promotional price against $4 / $20 for Claude Opus 5.5, and Microsoft’s own tests say it gives 96% of Opus 5.5’s performance at 5% of the cost.

Last updated:

Use it now

Chat

≈ 6 credits per reply · See all models

New accounts get 5 free credits once, valid for 7 days. They work on the lower-priced models here; the priciest ones need credits you bought.

Start chatting — 5 free credits

Per million tokens

Price per million tokens

checked 2026-11-19
-InputOutputOpenRouter id
MAI Code 2$0.5$2.5None
Claude Opus 5.5$4$20anthropic/claude-opus-5.5

The makers’ own facts

Side by side

-MAI Code 2Claude Opus 5.5
Released
Announced November 17, 2026 and available in GitHub Copilot that daySource: microsoft.ai · checked 2026-11-19
September 22, 2026Source: platform.claude.com · checked 2026-11-19
Price per 1M tokens (input / output)
$0.50 / $2.50 at a promotional price until May 17, 2027, then $1.00 / $5.00 (GitHub Copilot pricing table)Source: docs.github.com · checked 2026-11-19
$4 / $20Source: anthropic.com · checked 2026-11-19
Cached input per 1M tokens
$0.025 at the promotional price, then $0.05Source: docs.github.com · checked 2026-11-19
$0.20 (cache reads)Source: anthropic.com · checked 2026-11-19
One request, 10,000 tokens in and 1,000 out (our arithmetic)
About $0.0075 at the promotional price; about $0.015 from May 17, 2027Source: docs.github.com · checked 2026-11-19
About $0.06Source: anthropic.com · checked 2026-11-19
Where you can use it
GitHub Copilot Pro, Pro+, Max, Business and Enterprise, for eligible subscribers on annual plans, billed by usageSource: github.blog · checked 2026-11-19
Listed on OpenRouter as anthropic/claude-opus-5.5 at $4 / $20 per million tokensSource: openrouter.ai · checked 2026-11-19
Microsoft’s benchmark claim (average of DeepSWE 1.1, Terminal Bench 2.1, SWE-Atlas and FrontierCode v1.1)
96% of Opus 5.5’s performance (Microsoft’s own claim; all models at max reasoning effort)Source: microsoft.ai · checked 2026-11-19
100%, the baseline Microsoft compares againstSource: microsoft.ai · checked 2026-11-19
Cost in Microsoft’s comparison
5% of Opus 5.5’s cost, with MAI Code 2 at its 50% promotional priceSource: microsoft.ai · checked 2026-11-19
100%, the baseline Microsoft compares againstSource: microsoft.ai · checked 2026-11-19
Context window / max output
Not stated on Microsoft’s announcement or model pageSource: microsoft.ai · checked 2026-11-19
1M tokens / 128K tokensSource: platform.claude.com · checked 2026-11-19
How the maker describes the limits
"Explicitly not yet tuned for harder problems, like fundamental science research and very long-horizon agentic work"Source: microsoft.ai · checked 2026-11-19
"For long-running agentic coding and knowledge work"Source: platform.claude.com · checked 2026-11-19

Which one does the maker recommend?

  • Microsoft says MAI Code 2 and Code 2 Flash are "explicitly not yet tuned for harder problems, like fundamental science research and very long-horizon agentic work"; its own comparison with Opus 5.5 is an average over four everyday coding benchmarks.Source: microsoft.ai · checked 2026-11-19
  • Microsoft’s model page repeats "of Opus 5.5 performance at 5% of cost"; both pages are Microsoft’s own, and the footnote says performance and cost comparisons vary by workload, model configuration, reasoning effort and pricing assumptions.Source: microsoft.ai · checked 2026-11-19
  • Microsoft’s model page says MAI Code 2 Flash costs 1/5 of Claude Haiku 5.5 with 12% higher performance, and its blog compares Flash with GPT-6 Luna and Claude Haiku 5.5 across three coding benchmarks (Microsoft’s own figures, not tested by us).Source: microsoft.ai · checked 2026-11-19

Sample answers (AI-generated)

Same question, both models

Sample answers are pending: they have not been generated yet.

FAQ

Questions

Is MAI Code 2 better than Claude Opus 5.5?

Not by any independent measure we found. Microsoft’s own claim is that, averaged over DeepSWE 1.1, Terminal Bench 2.1, SWE-Atlas and FrontierCode v1.1, MAI Code 2 gives 96% of Opus 5.5’s performance, which means slightly below it, at a fraction of the cost. Microsoft also says MAI Code 2 is not yet tuned for harder problems or very long-horizon agentic work. We have not tested it.

How much cheaper is MAI Code 2 than Opus 5.5?

At list price, MAI Code 2 is one eighth of Opus 5.5 for now: $0.50 / $2.50 against $4 / $20 per million input / output tokens, until May 17, 2027, when GitHub’s table moves it to $1.00 / $5.00. Microsoft’s 5% figure is a different thing, a measured cost across its benchmark runs with MAI Code 2 at the promotional price.

Where can I use MAI Code 2?

In GitHub Copilot: Pro, Pro+, Max, Business and Enterprise plans for eligible annual subscribers, in VS Code, Visual Studio, Copilot CLI, JetBrains, Xcode and the other clients GitHub lists. Microsoft says MAI Code 2 Flash is in limited-access preview in Microsoft Foundry and that both models are coming "soon" to OpenRouter, Vercel, Hermes, OpenCode and Kilo; they were not on OpenRouter’s list when we checked on November 18, 2026.

What is MAI Code 2 Flash?

The cheaper sibling, announced the same day: $0.05 / $0.25 per million input / output tokens at the promotional price in GitHub’s table. Microsoft says it costs 1/5 of Claude Haiku 5.5 with 12% higher performance, and compares it with GPT-6 Luna as well. See Claude Haiku 5.5 and GPT-6 Luna for the models it is compared with.

When was MAI Code 2 released, and what came before it?

November 17, 2026, the day of Microsoft Ignite, in GitHub Copilot. Before it came MAI-Code-1-Flash, which reached Copilot in Visual Studio Code on June 2, 2026; GitHub says it will deprecate MAI-Code-1.1-Flash on December 18, 2026 and points users to the latest MAI Code Flash model.

Can I chat with MAI Code 2 and Opus 5.5 here?

Only with Opus 5.5: MAI Code 2 is not on our model provider. On this site a typical reply to Claude Opus 5.5 costs about 6 credits (a credit is about $0.01 in the $5 pack), billed by its real token cost, at least 1 credit. For the model one step above, see Claude Fable 5.2 vs Opus 5.5; Microsoft also compares MAI Code 2 with GPT-6.1 Sol.

Is this an official Microsoft or Anthropic page?

No. This page is unofficial and not affiliated with or endorsed by Microsoft or Anthropic. Every fact links to the page we read.