Microsoft has updated its coding model in a move that positions the company as a leading AI model developer capable of meeting enterprise business needs while remaining competitive in an increasingly crowded market.

On August 11, the tech giant released MAI-Code-1.1-Flash. According to Microsoft, the model produces more efficient code at a quarter of the cost compared to MAI-Code-1-Flash, which was first released in June.

The upgraded model focuses on CLI tasks — which involve using text or natural language to complete tasks — enabling enterprise developers to use fewer tokens to complete them, according to the company. CLI-based agent workflows have become a key testing ground for coding models, as they require models to interpret developer intent, generate functional code, and execute multi-step tasks reliably.

Previously, users paid $0.75 per million tokens for input and $4.50 per million tokens for output. Under the new pricing, users pay $0.20 per million tokens for input and $1.20 per million tokens for output on GitHub. For organizations deploying AI coding tools across large development teams, even modest per-token reductions compound into meaningful budget savings at scale.

Broader Industry Trend Toward Tokenomics

Microsoft's upgrade of MAI-Code is another instance in which an AI model provider is responding to enterprises' concerns about the high cost of using generative AI and building agentic workflows. Microsoft is not the first to cut prices soon after releasing a new model. Most recently, Google, Anthropic, and OpenAI have all repriced their models at lower price points in response to growing attention to tokenomics — the emerging practice of economizing AI token use.

Lower Pricing in a Competitive Market

For Microsoft, slashing prices could help it become more competitive in a landscape where Anthropic dominates as the coding model provider of choice for enterprises, alongside competition from OpenAI's Codex and SpaceX's Cursor. In recent months, Mistral and other AI vendors have also released models heavily focused on coding. Coding-specific models have emerged as a distinct and strategic category within the broader AI market, as enterprises increasingly view AI-assisted software development as one of the most measurable productivity gains from generative AI investment.

"For enterprises, they'll be looking at the lower token cost of [MAI]," said Lian Jye Su, an analyst at Omdia, a division of Informa TechTarget. "Right now, the token cost is through the roof. Any models that can help them to reduce that and yet deliver a high-quality coding experience or software accuracy are quite important."

Su added that it would be interesting to see how MAI-Code-1.1-Flash stacks up against SpaceXAI's Grok 4.5 in terms of performance, as that model also focuses on coding at a lower price point. Grok 4.5 is significantly more expensive, with input costing $2 per million tokens and output $6 per million tokens.

Beyond Pricing: Microsoft's Strategic Positioning

For Microsoft, the MAI update is about more than just pricing. It signals the company's intent to remain active in the AI market after ending its exclusivity agreement with model provider OpenAI earlier this year.

"It's about retaining users," Su said. He noted that with other major cloud providers — AWS and Google — fielding their own models, it comes as no surprise that Microsoft will continue to push its own models to enterprise developers.

"It's about keeping up with competition," he continued. For enterprises already integrated with Microsoft's broader product ecosystem, such as Copilot and GitHub, adopting any of the MAI models may represent a natural next step.

Source: AI Business