Microsoft Unveils Six New AI Models and Autopilot Agent at Ignite Conference
At its Ignite conference, Microsoft announced a suite of six new in-house MAI models, including MAI Code 2, MAI Image 2.7, and MAI Realtime 1, alongside the rollout of the Autopilot agent for Frontier customers. The company claims these models offer significant cost reductions compared to competitors like OpenAI and Anthropic, with MAI Code 2 achieving 96% of Claude Opus 5.5's performance at just 5% of the cost. This strategic move aims to reduce reliance on external providers while expanding enterprise productivity tools such as Copilot Home and Workspaces.
Editorial responsibility
- No named human review is recorded for this page.
- Reports are grouped by semantic similarity and deterministic rules. Language models may assist titles, summaries, translation and cross-source analysis; the page reads the event directly, while its address stays stable when the title changes.
- Summary covers the current reports
Cross-source coverage
Reporting timeline
Microsoft Unveils New AI Models and Autopilot Agent at Ignite
At Microsoft Ignite, the company announced a suite of new in-house MAI models, including MAI Code 2, MAI Image 2.7, and MAI Realtime 1, alongside the rollout of Autopilot, an always-on agent with memory for Frontier customers. The event highlighted Copilot Home, Workspaces, and Microsoft IQ to enhance enterprise productivity. A key strategic focus is efficiency, with Microsoft claiming MAI Code 2 achieves 96% of Claude Opus 5.5's performance at 5% of the cost through advanced orchestration like GitHub Fusion. While praising the ecosystem expansion and cost-cutting measures, the analyst notes that Autopilot currently offers less personalization than competitors like Grok Bot and remains a work in progress. The announcement signals Microsoft's push toward independence from external model providers by leveraging its own technology stack.
Read sourceMicrosoft Launches Six Low-Cost AI Models to Challenge OpenAI and Anthropic
On November 18, Microsoft announced the release of six new AI models covering programming, cybersecurity, image generation, and voice processing. According to Axios, these models aim to compete with industry leaders like OpenAI and Anthropic by offering significantly lower costs while maintaining high performance. Specifically, the MAI Code 2 model reportedly achieves 98% of GPT-6.1 Sol's performance at 40% of the cost, and 96% of Claude Opus 5.5's performance at just 5% of the cost. The MAI Cyber 2 Flash model surpassed GPT-6 Astra in a security test with task costs under 2% of the competitor's. Additionally, MAI Image 2.7 ranked first in image editing benchmarks. Satya Nadella's AI lead, Sudip Suleman, stated that Microsoft is accelerating the development of its own general-purpose AI models to reduce reliance on external providers. The company also plans to open-source model weights, enabling enterprises to customize models using their own data.
Microsoft Launches MAI Code 2 Flash on GitHub Copilot and Announces New AI Models
Microsoft announced a suite of new AI model updates at its Ignite conference, including the release of MAI Code 2 Flash on GitHub Copilot with limited access preview in Foundry. The company also introduced MAI Cyber 2 Flash for upcoming security products and MAI Image 2.7 across multiple platforms like Vercel and OpenRouter. Additionally, MAI Code 2 was launched on GitHub Copilot, while MAI Transcribe 2 is now available in Dragon Copilot for medical scenarios. Performance claims highlight that AT&T testing showed MAI Code 2 Flash outperforming GPT-6.1 Sol at 50 times lower cost. Furthermore, the voice AI model MAI Realtime 1 reportedly surpassed GPT Realtime 2.1 in public benchmarks and internal evaluations. These releases mark significant expansions in Microsoft's AI capabilities across coding, cybersecurity, image generation, and voice processing.
Read source