Microsoft

Microsoft Research has released Fara-7B, a 7-billion parameter small language model designed for computer automation tasks. The model, which took 2.5 days to train on 64 H100 GPUs, can navigate websites to complete tasks like booking restaurants and shopping, using screenshots as input with a 128K token context window.

Microsoft is removing Claude Code access from its Experiences + Devices division by June 30, 2026, redirecting thousands of engineers to GitHub Copilot CLI instead. The decision follows six months of Claude Code proving more popular than Microsoft's own coding tool among internal developers.

Microsoft Edge mobile version 148 and higher integrates six AI-powered features from its desktop version, including the ability to summarize multiple tabs simultaneously, generate podcasts from web pages, and recall browsing history for continued conversations. The update also adds a Journeys feature that tracks research topics and a Study and Learn mode for interactive quizzes.

Microsoft is updating Edge to let Copilot read and analyze content across all open browser tabs simultaneously. The update includes AI-generated podcasts from tabs, study mode with quizzes, and long-term conversation memory.

Microsoft CEO Satya Nadella disclosed that M365 Copilot has reached 20 million paid enterprise seats during the company's quarterly earnings call. Weekly engagement now matches Outlook usage levels, with queries per user up 20% quarter-over-quarter.

Microsoft's GitHub is ending flat-rate billing for Copilot on June 1, 2026, switching to usage-based metered tokens after acknowledging the request-based model is no longer sustainable. Copilot Pro subscribers ($10/month) will receive 1,000 GitHub AI Credits monthly, with each credit worth $0.01.

Microsoft has pushed agentic Copilot features into general availability across Word, Excel, and PowerPoint. The AI assistant can now make direct edits to documents, spreadsheets, and presentations rather than just suggesting changes from a sidebar.

Microsoft is rolling out Agent Mode as the default Copilot experience in Word, Excel, and PowerPoint this week for all Microsoft 365 Copilot and Premium subscribers. The feature, previously called 'vibe working,' allows the AI to execute multi-step edits directly in documents with real-time visibility into each action.

GitHub has paused new subscriptions for Copilot Pro, Pro+, and Student plans due to compute capacity constraints. The company cites agentic workflows as consuming significantly more resources than its original pricing structure anticipated, forcing tighter usage limits and a shift away from flat-rate billing.

GitHub Copilot customers are experiencing rate limits lasting up to 181 hours after the company fixed a token counting bug that had been undercharging for usage of newer models including Claude Opus 4.6 and GPT-5.4. GitHub has suspended Pro free trials and removed Anthropic's Opus 4.6 Fast model from Pro+ tiers.

Microsoft has added track changes monitoring, comment management, and table of contents insertion to Copilot in Word. The features run on Microsoft's "Work IQ" adaptation layer and are currently available only through the Office Insiders Beta Channel Frontier program on Windows.

Microsoft is testing OpenClaw-like features for Microsoft 365 Copilot aimed at enterprise customers, the company confirmed to The Information. The agent would run continuously to complete multi-step tasks over extended periods, distinguishing it from Microsoft's existing cloud-based agents like Copilot Cowork and Copilot Tasks.

Microsoft is testing OpenClaw-style autonomous agents for 365 Copilot that would run continuously to complete tasks on behalf of users, according to The Information. The company plans to demonstrate some features at its Build conference on June 2nd.

Microsoft has begun removing Copilot branding and buttons from Windows 11 applications including Notepad and Snipping Tool, replacing them with generic icons like a pen symbol. The underlying AI-powered features remain functional but are now labeled as "writing tools" rather than Copilot. This follows user complaints about forced Copilot integration and inconsistent experiences across apps.

Microsoft is removing Copilot buttons from Windows 11 apps including Notepad, Snipping Tool, Photos, and Widgets as part of a broader effort to reduce "unnecessary Copilot entry points." The underlying AI features remain intact, with Notepad's Copilot button replaced by a "writing tools" menu that retains the same AI-powered functionality.

Microsoft's Bing team has open-sourced Harrier, a 27-billion-parameter embedding model that supports over 100 languages and features a 131,072-token context window. The model ranks first on the MTEB v2 multilingual benchmark, outperforming proprietary offerings from OpenAI and Amazon, and is available on Hugging Face under the MIT license.

Microsoft released public preview versions of three proprietary AI models: MAI-Transcribe-1 for speech recognition across 25 languages at 50% lower GPU cost than alternatives, MAI-Voice-1 for speech synthesis generating 60 seconds of audio in under a second, and MAI-Image-2 for text-to-image generation. The models are available exclusively through Microsoft Azure AI Foundry and already power Copilot, Bing, and PowerPoint.

Microsoft AI released three foundational models on April 2: MAI-Transcribe-1 for speech-to-text across 25 languages, MAI-Voice-1 for audio generation, and MAI-Image-2 for video generation. The company positions these models as cheaper alternatives to Google and OpenAI offerings. Models are available on Microsoft Foundry with pricing starting at $0.36 per hour for transcription.

Microsoft has released MAI-Transcribe-1, a speech-to-text model that achieves the lowest word error rate on the FLEURS benchmark across 25 languages, outperforming Whisper-large-V3, GPT-Transcribe, and Gemini 3.1 Flash-Lite. The model runs 2.5 times faster than Microsoft's previous Azure Fast offering and costs $0.36 per audio hour.

Microsoft released the Harrier-OSS embedding model family, comprising three variants with 270M, 600M, and 27B parameters. The largest model achieves 74.3 on the Multilingual MTEB v2 benchmark. All models support 32,768 max tokens and multilingual inputs across 40+ languages.