Microsoft cut prices for its coding model by 75% as competition intensifies with Anthropic and OpenAI for enterprise developers.

The company released MAI-Code-1.1-Flash on August 11, pricing input tokens at $0.20 per million and output at $1.20 per million. The previous version cost $0.75 for input and $4.50 for output tokens.

Microsoft said the upgraded model completes CLI tasks more efficiently using fewer tokens than its June predecessor. The model focuses on command-line interface tasks that use natural language to complete coding workflows.

Pricing War Heats Up

The price cut follows similar moves by Google, Anthropic and OpenAI as enterprises scrutinize AI token costs. Anthropic currently dominates enterprise coding with Claude, while competitors like Cursor and Mistral have launched coding-focused models.

"For enterprises, they'll be looking at the lower token cost," said Lian Jye Su, an analyst at Omdia. "Right now, the token cost is through the roof."

Microsoft's pricing undercuts xAI's Grok 4.5, which charges $2 per million input tokens and $6 for output. The move comes after Microsoft ended its exclusivity agreement with OpenAI earlier this year.

"It's about retaining users," Su said. "With the other major cloud providers, AWS and Google, fielding their own models, it's no surprise that Microsoft will continue to push its own models."

Enterprises already using Microsoft's Copilot and GitHub may find MAI models a natural extension of their existing workflows. The company positions the update as part of staying competitive in the rapidly evolving AI coding market.

Microsoft plans to continue developing its MAI model family as it competes for enterprise AI workloads beyond its OpenAI partnership.