# MAI-Code-1-Flash

Microsoft, released 2026-06-02

**Verdict: conditional.** MAI-Code-1-Flash was Microsoft's first in-house coding model, a sparse MoE priced at $0.75/$4.50 per 1M tokens. GitHub retired it from Copilot on 10 September 2026 and names MAI-Code-1.1-Flash, at $0.20/$1.20, as the replacement; it no longer appears in Copilot's supported-models or pricing tables. Copilot users should switch to 1.1-Flash, which is cheaper and scores higher on Terminal-Bench 2.1. Last verified 2026-06-28.

## Pricing

- Input: $0.75 / 1M tokens
- Output: $4.50 / 1M tokens
- Context: Retired from Copilot Sep 10 2026; 256K ctx

## Strengths

- Historical comparison only: GitHub retired it from Copilot on 10 September 2026, so new Copilot work should use MAI-Code-1.1-Flash
- Fast, iterative agentic coding workflows where latency and token cost matter more than frontier reasoning, on any host that still serves it
- New Copilot workflows: the model no longer appears in GitHub Copilot's supported-models or pricing tables after its 10 September 2026 retirement
- Cost-sensitive teams: its successor MAI-Code-1.1-Flash lists at $0.20/$1.20 per 1M against this model's $0.75/$4.50
- Complex multi-file architectural changes or novel algorithm design requiring frontier-level code reasoning
- General-purpose non-coding tasks — this is a specialized coding model, not a general assistant

## Benchmarks

- SWE-Bench Pro: 51.2% (vs Haiku 4.5: 35.2%) (https://microsoft.ai/news/introducingmai-code-1-flash/)
- SWE-Bench Verified: Beats Haiku 4.5 with up to 60% fewer tokens (https://microsoft.ai/news/introducingmai-code-1-flash/)
- AIME 2026: 92.5% (vs Haiku 4.5: 83.3%) (https://microsoft.ai/pdf/MAI-Code-1-Flash-Model-Card.PDF)
- GPQA Diamond: 84.6% (vs Haiku 4.5: 73.2%) (https://microsoft.ai/pdf/MAI-Code-1-Flash-Model-Card.PDF)
- IF Bench: 75.0 (vs Haiku 4.5: 46.1) (https://microsoft.ai/pdf/MAI-Code-1-Flash-Model-Card.PDF)
- Adversarial reasoning: 85.8% adjusted accuracy (https://microsoft.ai/news/introducingmai-code-1-flash/)
- Frontier Math (Tier 1-3): 6.3% (vs Haiku 4.5: 2.8%) (https://microsoft.ai/pdf/MAI-Code-1-Flash-Model-Card.PDF)
- Frontier Science: 58.2% (vs Haiku 4.5: 42.3%) (https://microsoft.ai/pdf/MAI-Code-1-Flash-Model-Card.PDF)
- HLE: 18.0% (vs Haiku 4.5: 9.5%) (https://microsoft.ai/pdf/MAI-Code-1-Flash-Model-Card.PDF)

## Sources

- [GitHub Copilot models and pricing: MAI-Code-1-Flash](https://docs.github.com/copilot/reference/copilot-billing/models-and-pricing)
- [Microsoft AI — Introducing MAI-Code-1-Flash (official blog)](https://microsoft.ai/news/introducingmai-code-1-flash/)
- [GitHub Copilot: Supported AI models](https://docs.github.com/copilot/reference/ai-models/supported-models)
- [Microsoft AI — MAI-Code-1-Flash Model Card (PDF)](https://microsoft.ai/pdf/MAI-Code-1-Flash-Model-Card.PDF)
- [GitHub Changelog](https://github.blog/changelog/2026-06-26-mai-code-1-flash-for-copilot-business-and-copilot-enterprise/)
- [GitHub Changelog: Upcoming deprecation of MAI-Code-1-Flash (retirement 2026-09-10)](https://github.blog/changelog/2026-08-11-upcoming-deprecation-of-mai-code-1-flash/)

Canonical: https://neomanex.com/models/mai-code-1-flash