Microsoft (MSFT.US) Ignite Conference Releases Six Dedicated AI Models: Challenging OpenAI and Anthropic with Lower Costs
Microsoft is making significant progress in building its own cutting-edge general-purpose AI models and has launched six dedicated AI models at the Ignite conference.
Zhitong Finance APP reports that Microsoft (MSFT.US) is accelerating its in-house AI model strategy. Mustafa Suleyman, head of AI at Microsoft, stated that the company is making significant progress in building its own cutting-edge general AI models. At the Ignite conference, Microsoft launched six specialized AI models, claiming these models can achieve comparable performance to leading products from OpenAI and Anthropic at substantially lower costs. According to Microsoft, tasks that previously took weeks to complete can now be accomplished in just minutes.
Key Takeaway: Transition from Relying on External Models to In-House Competition
For a long time, Microsoft’s most advanced models depended on OpenAI and other AI companies. This release shows that Microsoft can develop competitive proprietary technology and signals its intent to compete aggressively on price. On Tuesday, Microsoft unveiled six new AI models at the Ignite conference in San Francisco, covering coding, cybersecurity, image generation, and speech.
Specifically:
MAI Code 2: Microsoft says this coding model achieves 98% of GPT-6.1 Sol’s performance at 40% of its price; it reaches 96% of Anthropic Claude Opus 5.5’s performance at just 5% of its cost. This flagship coding model is designed to handle complex, time-consuming programming tasks, work across multiple files, and comprehend entire software projects.
MAI Code 2 Flash: A smaller, faster model. Microsoft claims it matches GPT-6 Luna’s performance at 65% of its price; it matches Claude Haiku 5.5 at 20% of its cost.
MAI Cyber 2 Flash: A cybersecurity model. Microsoft states it outperforms GPT-6 Astra in selected security benchmarks, with a per-task cost of less than 2% compared to the latter. Microsoft also plans to provide the cybersecurity-specific model to a limited client group to help address the surge in software vulnerabilities exposed by AI tools.
Other models: According to Microsoft, MAI Image 2.7 ranks first in image editing; new transcription and real-time speech models aim to challenge leading competitors.
Strategy: Specialized Models for Cost Advantage
Cost pressure on enterprise clients is a major background factor for this release. Suleyman noted he constantly hears complaints from enterprises that “token spending is skyrocketing.” He said, “Our commitment is to offer the best value in the market, enabling you to truly scale agents; we also know you want models that truly belong to you.”
Microsoft’s strategy is to build specialized models for specific tasks, thereby achieving comparable performance at significantly lower costs. Suleyman stated, “This basically means you can pack more tokens, knowledge, and expertise into the same parameter count.” He also said, “This is not just a single model or a flash in the pan. It’s a system, a continually advancing machine we’re building to optimize both performance and price together.”
However, this strategy involves trade-offs: models optimized for everyday coding tasks may not perform as well in advanced mathematics or scientific research. Additionally, Microsoft’s performance comparisons depend on selected benchmarks, and some of the price advantages reflect promotional discounts.
Enterprise Platform and Cost Control
Beyond the models themselves, Microsoft also emphasized enterprise-grade platforms and cost control tools at Ignite. Executives highlighted the Microsoft IQ platform, aimed at helping enterprises organize and share employee knowledge, expertise, and tools, allowing AI systems to learn how enterprises operate and continue to build upon this foundation.
Microsoft CEO Satya Nadella emphasized that the new AI models cover scenarios such as coding, cybersecurity, and decision-making. He also highlighted Copilot Fusion, saying it can orchestrate multiple models to deliver cutting-edge results at significantly lower costs. Microsoft also showcased a model selection feature, enabling the platform to choose suitable models for each task—maximizing output quality while managing per-request costs.
Industry Context: Intensifying Competition in Cutting-Edge AI
As Microsoft targets industry leaders, frontier AI itself is also evolving rapidly. OpenAI on Tuesday released GPT-6.2 Astra and GPT-6.2 Sol, with the new Astra priced 20% lower than the previous version. At the same time, Microsoft is advancing a cutting-edge general large language model, aiming to compete across a broader range of tasks.
Suleyman stated that Microsoft has achieved state-of-the-art pretraining results and is making strong progress in reinforcement learning, but he refused to provide a release timeline for this model. He said, “My team is building true superintelligence.”
Potential Impact: Reshaping Microsoft’s Relationship with OpenAI
The new models could reshape Microsoft's relationships with OpenAI and other external AI providers. MAI Code 2 will be available in GitHub Copilot, potentially shifting some coding workloads from external models to Microsoft’s own. When asked what share of Copilot traffic Microsoft expects to shift to its own models, Suleyman declined to provide a target but predicted significant adoption.
For enterprise business, Microsoft also offers clients licensed model weights so they can customize models with proprietary data, but with the requirement that deployment be within Azure. Suleyman said this approach “feels like open source,” while also providing greater security and control.
Security and Alignment Listed as Priorities
Suleyman said that recent cybersecurity incidents—such as the OpenAI AI agent breach of the Hugging Face platform—have prompted him to focus equally on critical safeguards as well as performance. Preventing AI from taking harmful actions and ensuring it behaves as intended, i.e., “alignment” research, are his top priorities.
At the opening of the conference, Nadella also acknowledged last week’s cross-region service outage, stating the company takes customer trust seriously, has work to do, and is committed to ongoing learning and improvement.
Going forward, the market will focus on two aspects: first, how quickly Microsoft can translate advances in specialized AI into competitive general models; and second, whether its cost advantages are sufficient to prompt clients to switch from established competitors to Microsoft.
On Tuesday, Microsoft shares closed down 2.27%, up 5.3% year-to-date—underperforming the S&P 500. As of press time, the stock was up 0.9% in after-hours trading.
Disclaimer: The content of this article solely reflects the author's opinion and does not represent the platform in any capacity. This article is not intended to serve as a reference for making investment decisions.
You may also like
NZD/USD Price Forecast: Declines below 0.5500, bearish bias intact below 100-day SMA

Anthropic IPO may deal a heavy blow to Nvidia (NVDA.US) short sellers! Cramer issues a warning, Bezos discusses Jensen Huang's successor
Anthropic will be listed on Nasdaq tonight, and Nvidia, as both investor and supplier, is receiving attention. Cramer warns short sellers.
"AI Central Bank" initiates quantitative tightening! Nvidia's new CEO tightens investment gates, Neocloud's funding lifeline under threat
Nvidia's interim CEO Kress announced during the financial results meeting that the company will strictly evaluate investments and guarantees, ending the "broad support" approach of Jensen Huang's era. Constrained by supply chain commitments as high as $390 billion, the company will tighten financial backing for the highly leveraged Neocloud sector.

Nike (NKE.US) Announces "Full Speed Ahead" Strategy: Five-Year Revenue Growth Target of $5 Billion, Professional Sports Business Becomes Growth Engine
Nike officially launched the "Full Stride Ahead" strategy at its 2026 Investor Day and announced its financial framework for fiscal year 2031. The company plans to use fiscal year 2027 as the baseline, aiming to achieve approximately $5 billion in additional revenue by fiscal year 2031, an operating profit margin of 12% to 14%, and adjusted earnings per share rising to $3 to $3.50.
