Microsoft’s new coding model promises near-Opus performance at 5% of the cost

The New Stack (AI)
Read full story· 3 min
Microsoft unveiled two new coding AI models, Code 2 and Code 2 Flash, at Ignite 2026, promising near-top-tier performance at significantly reduced costs compared to GPT-6.1 Sol and Claude Opus 5.5. These models are optimized for enterprise coding tasks and visual coding, focusing on efficiency over broad generalist capabilities.

What happened

On Tuesday at its Ignite conference in San Francisco, Microsoft introduced two new coding-focused AI models, Code 2 and Code 2 Flash. These models are designed for enterprise coding and visual coding tasks, with an emphasis on efficiency and cost savings. Microsoft stated that Code 2 achieves nearly the same performance as leading models like GPT-6.1 Sol and Claude Opus 5.5 but at a much lower cost. Both models are now available in GitHub Copilot, with Code 2 Flash also in limited preview on Microsoft Foundry.

Key details

  • Code 2 delivers 98% of GPT-6.1 Sol's performance at 40% of the cost.
  • Code 2 matches 96% of Claude Opus 5.5's performance at just 5% of the cost.
  • Code 2 Flash matches Claude Haiku 5.5 and GPT-6 Luna on key coding benchmarks at 20% and 65% of the cost, respectively.
  • The models are optimized for enterprise coding and visual dashboard creation, not for broader or more complex tasks.
  • Support for Code 2 and Code 2 Flash is coming soon to OpenRouter, Vercel, Hermes, Open Code, and Kilo.

Why it matters

Enterprise developers and companies using coding assistants now have access to models that offer nearly top-tier performance for a fraction of the usual cost. This can reduce expenses for organizations that rely on AI-powered coding tools, making advanced capabilities more accessible. The focus on specialized coding tasks means teams can expect efficient solutions tailored to their daily workflows, rather than paying for broader generalist models they may not need.


More in Dev