Hot Take Nov 17, 2026 · 1 min read

Microsoft says its new coding model, MAI Code 2, hits 96% of Claude Opus 5.5's performance at 5% of the price.

Microsoft says its new coding model, MAI Code 2, hits 96% of Claude Opus 5.5's performance at 5% of the price. That's the company's own benchmark claim, and it's the number I expect engineering leaders to be passing around this week.

At Ignite 2026 in San Francisco today, Microsoft announced seven new AI models, framed by outlets like CNBC-TV18 and Neowin as a direct move against OpenAI and Anthropic. Per Microsoft, MAI Code 2 reaches 98% of GPT-6.1 Sol's performance at about 40% of the price across four benchmarks, and developers preferred it over GPT-6 Sol 64% of the time in a blind frontend test.

On the security side, Microsoft says MAI Cyber 2 Flash matches GPT-6 Astra's quality at under 2% of the cost per task on CyberGymE2E. And Decision 1.1, its router that sends requests to bigger or cheaper models as needed, is already used by 4,000 customers, with Microsoft claiming it's 50x faster and 79x cheaper than GPT-Sol 6.1.

These are all vendor numbers, so I'd wait for independent testing before rewriting any procurement plans. But the direction is hard to miss. If the price gap holds up even partially, I'm curious how long OpenAI stays the default model inside Microsoft's own products. $MSFT

Sources:
Microsoft AI: https://microsoft.ai/news/microsoft-ignite-2026/
CNBC-TV18: https://www.cnbctv18.com/technology/microsoft-takes-on-openai-and-anthropic-with-new-ai-models-for-coding-and-cybersecurity-20036327.htm
Neowin: https://www.neowin.net/news/microsoft-takes-on-openai-and-anthropic-with-seven-new-ai-models-at-ignite-2026/

Not financial advice. For informational and educational purposes only.

Track your live portfolio alongside our daily institutional research

Connect your brokerage in 60 seconds and get full access to Position Tracker Substack research inside your dashboard.

Start your free trial →