Microsoft adds MAI-Code-1.1-Flash to GitHub Copilot with 73% lower

MAI-Code-1.1-Flash, Microsoft’s newest small‑tier coding model, is now live in GitHub Copilot with a list price 73% lower than its predecessor, MAI‑Code‑1‑Flash, aiming to cut costs while adding native image‑understanding capabilities【1】. The move targets developers seeking a balance of capability and affordability, especially in lightweight coding workflows.

Model upgrades and efficiency gains

MAI‑Code‑1.1‑Flash builds on the earlier MAI‑Code‑1‑Flash by adding native vision support, which lets the model interpret images as part of coding tasks. Microsoft also reports a 25% improvement in token efficiency—meaning the model streams tokens 25% faster and uses 25% fewer tokens per task—translating into quicker responses and lower per‑request costs【2】. Benchmarks show a 22% lift on Terminal‑Bench 2.1 in the Copilot CLI and a 15% boost on .NET tasks, while code survival rose 4% and return visits grew 9%【2】.

Pricing and rollout strategy

The 73% lower list price is reflected in a 0.25× premium request multiplier for annual Copilot subscribers, positioning the model as a cost‑effective option for “lightweight coding workflows”【1】. Access is automatic for free and student users, while Pro, Pro+, Max, Business, and Enterprise plans can manually select the model via the Copilot model picker across a wide range of IDEs, including VS Code, Visual Studio, JetBrains, Eclipse, and Xcode【1】. Enterprise and Business administrators must enable the model policy, which is off by default, giving Microsoft control over rollout pacing.

Competitive context

Microsoft’s emphasis on token efficiency mirrors industry pressure to reduce inference costs. While prior claims for MAI‑Code‑1‑Flash highlighted up to 60% fewer tokens on SWE‑Bench Verified, the new 25% token‑efficiency figure provides a concrete, measurable improvement over the June‑2026 launch model【3】. Competitors such as Anthropic’s Haiku and other lightweight coding models have not disclosed comparable pricing, making Microsoft’s public cost reduction a distinctive market signal.

What to watch

• Policy enablement: Adoption will depend on Enterprise and Business admins turning on the MAI‑Code‑1.1‑Flash policy.
• Performance data: Real‑world token‑usage and latency metrics from Copilot users will reveal whether the 25% efficiency claim holds at scale.
• Rival responses: Expect competing AI‑coding providers to announce pricing or efficiency upgrades to retain enterprise customers.

Microsoft’s rollout of MAI‑Code‑1.1‑Flash underscores a strategic push to pair higher coding quality with lower operating costs, potentially reshaping pricing expectations for AI‑assisted development tools. The real test will be whether developers adopt the model at scale and whether rivals follow suit with comparable cost‑efficiency claims.