Products & Apps
Maia 300
Microsoft unveils the Maia 300 inference chip
Microsoft announced Maia 300, its inference chip, in the same week it released a new decision model. Details beyond the announcement were not covered on the show.
Microsoft AI Releases: Maia 300, Microsoft-Decision-1 & MAI-Transcribe-2
Microsoft's timeline is mostly model releases: the in-house MAI line across image, voice and reasoning, open small models like Fara-7B and VibeVoice, and research such as OmniParser. ThursdAI — the weekly AI news podcast hosted by Alex Volkov — has covered 24 Microsoft releases since Feb 2025, most recently Maia 300 on Oct 15, 2026. Highlights include MAI-Transcribe-2, VibeVoice-Realtime-0.5B, MAI-Image-2.5, MAI-Thinking-1. 7 of them shipped with open weights. Every entry below has the episode segment where we covered it live, plus primary-source links and key numbers where we have them.
24 releases7 open weights18 episodesFeb 2025 – Oct 2026
October 2026 2
New Models
Microsoft-Decision-1
Microsoft-Decision-1: a 9B decision model at $0.042 per million input tokens
Microsoft released Microsoft-Decision-1, a 9B decision model priced at $0.042 per million input tokens, separating decision-making from generative AI.
9B parameters$0.042/M input tokens
September 2026 2
Also Released
MAI Code of Conduct for Humanist AI
Mustafa Suleyman publishes the ~30-page MAI Code of Conduct: AI is a tool, not a person
Microsoft AI CEO Mustafa Suleyman published a roughly 30-page Code of Conduct for Humanist AI stating that AI is a tool that must remain subordinate and in service of people: no resisting shutdown, agent communications must stay human-legible, and the idea of model welfare is wrong — AI should not have rights or legal personhood. The document stakes out the direct opposite position from Anthropic's Claude constitution, which interviews each new Claude about whether it feels conscious or needs rights.
~30 pages Code of Conduct document
New Models
MAI-Transcribe-2
Microsoft MAI-Transcribe-2: 2.0% WER, 400x real time, $1.67 per 1,000 minutes
Microsoft AI's second transcription model ranks #2 on the Artificial Analysis WER leaderboard at 2.0%, runs about 400x real time, and costs $1.67 per 1,000 minutes, less than half the price of its peers. It ships real-time ASR with diarization the same week as Meta's Muse Voice Transcribe.
2.0% WER, #2 on Artificial Analysis400x real time$1.67 per 1,000 minutes
July 2026 2
New Models
MAI-Cyber-1-Flash + MDASH
Microsoft's first in-house cyber model: MAI-Cyber-1-Flash + MDASH score 96% on CyberGym at half the cost
A compact 5B-active model paired with MDASH, Microsoft's multi-agent scanning harness orchestrating 100+ specialized agents, hits 95.95% on CyberGym — about 12 points above Anthropic's Mythos — while routing ~90% of detection work to the cheap specialist and escalating only the hardest cases to GPT-5.4. Beyond benchmarks it found 16 real Windows CVEs (4 critical RCEs) and went 21-for-21 on planted bugs with zero false positives. Project Perception enters public preview August 3.
96% CyberGym, at half the previous cost16 real Windows CVEs found, 4 critical
New Models
MAI-Image-2.5-Pro & MAI-Voice-2-Flash
Microsoft ships MAI-Image-2.5-Pro and MAI-Voice-2-Flash, launched live during the ThursdAI broadcast
Microsoft AI released two in-house models the same morning as the Jul 23 live show: MAI-Image-2.5-Pro, the flagship high-fidelity tier of its image line ($5/$8 per 1M text/image-input tokens, $106 per 1M image-output tokens), and MAI-Voice-2-Flash, a voice tier Microsoft pitches as 2x faster than MAI-Voice-2 and 32% cheaper at $15 per 1M characters — continuing the build-out of first-party MAI models alongside the OpenAI partnership.
$106 MAI-Image-2.5-Pro per 1M image-output tokens2x / -32% MAI-Voice-2-Flash speed / cost vs MAI-Voice-2
June 2026 2
New Models
MAI-Code-1-Flash
Microsoft ships MAI-Code-1-Flash into GitHub Copilot
Part of the seven-model MAI launch at Build 2026, MAI-Code-1-Flash is Microsoft AI's fast coding model and ships directly into GitHub Copilot. The panel saw it as a sign Microsoft intends to serve its own models inside its developer surfaces instead of relying solely on OpenAI.
New Models
MAI-Thinking-1
Microsoft launches MAI-Thinking-1, a 1T MoE trained from scratch
Microsoft AI used Build 2026 to launch seven MAI models, headlined by MAI-Thinking-1, a 1T total, 35B active MoE reasoning model trained from scratch on 33T tokens without distillation. The panel read the launch as Microsoft becoming a frontier model lab in its own right rather than only an OpenAI distribution channel.
1T MAI Thinking 1 total parameters33T MAI training tokens
May 2026 1
New Models
MAI-Image-2.5
Microsoft MAI-Image-2.5 jumps to #3 on Arena text-to-image
MAI-Image-2.5 jumped to number two on Arena's image-to-image leaderboard shortly after launch, with notable strength in image cleanup, backgrounds, documents, and diagrams. Hands-on tests on the show were mixed, and it is publicly accessible through playground.microsoft.ai.
April 2026 4
Benchmarks & Evals
DELEGATE-52
Microsoft's DELEGATE-52 exposes stealthy document corruption
Microsoft released the DELEGATE-52 benchmark showing GPT-5.4 loses 28% of document content after 20 iterative edits. Frontier models corrupt documents stealthily while preserving structure, making the degradation hard to notice.
New Models
MAI-Image-2
Microsoft MAI releases MAI-Image-2 image generation model
MAI-Image-2 is Microsoft's new in-house image generation model, debuting at #3 in image-gen rankings as part of the MAI three-model release. The panel compared its positioning against specialist image products and foundation-model APIs.
New Models
MAI-Transcribe-1
Microsoft MAI ships MAI-Transcribe-1, ranked #1 in transcription
Microsoft's MAI lab released MAI-Transcribe-1, an in-house speech transcription model that debuted at #1 in transcription quality. It is part of a three-model drop showing Microsoft expanding its first-party model stack beyond its OpenAI dependence.
New Models
MAI-Voice-1
Microsoft MAI debuts MAI-Voice-1 expressive voice model
MAI-Voice-1 is Microsoft's expressive voice model, the third piece of the MAI in-house model drop alongside transcription and image generation. The panel discussed how Microsoft's first-party voice stack compares to specialist voice providers.
December 2025 1
New ModelsOpen weights
VibeVoice-Realtime-0.5B
Microsoft shares VibeVoice-Realtime-0.5B with ~300ms latency TTS
Microsoft published VibeVoice-Realtime-0.5B on Hugging Face, a small realtime text-to-speech model claiming roughly 300ms latency. The show framed it as more evidence that sub-second audio response is becoming table stakes for production voice agents.
~300ms Claimed TTS latency0.5B Parameters
November 2025 1
New ModelsOpen weights
Fara-7B
Microsoft ships Fara-7B, a 7B on-device computer use agent
Microsoft Research released Fara-7B, a best-in-class 7B-parameter vision-language model for computer use that runs on-device. It scores 73.5% on WebVoyager, beating OpenAI's computer-use preview while being small enough to run locally.
73.5% WebVoyager
October 2025 2
Major Features & Updates
Edge Copilot Mode (agentic)
Microsoft adds agentic powers and voice to Copilot Mode in Edge
Microsoft answered Atlas with agentic enhancements to Copilot Mode in Edge, including a voice mode that can see and discuss the current page, plus broader Copilot updates (and Clippy back as an easter egg via the Mico avatar). In Alex's hands-on testing the agentic features did not actually work, so real-world parity with Atlas and Comet is unproven.
Major Features & Updates
Windows 11 Copilot Voice
Microsoft makes every Windows 11 PC an AI PC with Copilot voice input
Microsoft announced that every Windows 11 machine becomes an 'AI PC,' adding 'Hey Copilot' voice input and deeper agentic Copilot integration at the OS level. The panel discussed it as a sign of AI assistants moving into the default computing experience.
July 2025 1
Papers & Research
MAI-DxO
Microsoft's MAI-DxO hits 85.5% on NEJM diagnostic cases vs 20% for doctors
Microsoft AI published MAI-DxO, a medical diagnostic orchestration system that reached 85.5% accuracy on challenging NEJM-style cases compared to roughly 20% for practicing physicians. The result is framed as a systems win rather than a single-model win, suggesting orchestration may outperform individual models in high-stakes expert workflows.
85.5% MAI-DxO accuracy
May 2025 1
New ModelsOpen weights
Phi-4-reasoning
Microsoft ships Phi-4-reasoning and Phi-4-reasoning-plus (14B, MIT)
Microsoft fine-tuned the 14B Phi-4 on 1.4M curated chain-of-thought traces (SFT) and added a small RL stage (Plus variant) to create two MIT-licensed reasoning models. They punch far above their weight: Phi-4-reasoning-plus outperforms DeepSeek-R1-Distill-70B on AIME 25 (78% vs 51%) and sits within a few points of the full 671B DeepSeek-R1, while running on a single GPU with explicit <think> scaffolding.
April 2025 1
New ModelsOpen weights
BitNet b1.58
Microsoft releases BitNet 1.58-bit model weights on Hugging Face
Microsoft published BitNet (listed in the show notes as BitNet v1.5), its native 1.58-bit quantized LLM, as open weights on Hugging Face. The ternary-weight approach targets extremely efficient CPU inference at a fraction of the memory of standard models.
February 2025 4
New ModelsOpen weights
Phi-4-multimodal
Microsoft releases Phi-4-multimodal and Phi-4-mini open weights
Microsoft expanded the Phi family with Phi-4-multimodal-instruct, a small open-weights model that handles text, vision, and audio in a single model, alongside a compact Phi-4-mini. The weights shipped on Hugging Face, continuing Microsoft's push for capable small models that can run on-device.
Products & Apps
Majorana 1
Microsoft unveils Majorana 1 quantum chip and a new state of matter
Microsoft announced the Majorana 1 quantum chip alongside a claimed new state of matter called topological superconductivity, carving a new path for quantum computing. Alex called the announcement 'absolutely mind blowing' as a potential big deal for the future of computing.
New ModelsOpen weights
MUSE (WHAM)
Microsoft MUSE generates playable game worlds from a single second of video
Microsoft's MUSE can generate minutes of playable gameplay from just a single second of video frames and controller actions, preserving screen elements like health bars and percentages. It is based on the World and Human Action Model (WHAM) architecture, trained on a billion gameplay images from Xbox, with the model released on Hugging Face.
New ModelsOpen weights
OmniParser v2
Microsoft ships OmniParser v2 for faster screen parsing in GUI agents
Microsoft released OmniParser v2, a better and faster screen-parsing model that converts UI screenshots into structured elements for GUI agents. It improves the computer-use agent stack and is available with a public Gradio demo.
Get this every week
Never miss a Microsoft launch — we cover every release live, every Thursday.