What is happening in AI
New models, useful tools and ideas worth noticing. The facts first, then what they mean.
Le Figaro journalism is coming to ChatGPT
A publisher partnership puts attribution and original reporting in focus.
GPT-6.2: a new focus on getting work done
OpenAI introduces Astra and Sol for professional tasks and everyday work.
Gemini moves from conversation to errands
Google expands Tasks with calling, scheduling and other everyday workflows.
ChatGPT gets a dedicated sales toolkit
A Sales plugin connects customer research, preparation and follow-up workflows.
Strengthening NATO and allied cyber defenses with AI
A new item from OpenAI. Follow the source link to read the complete publication.
Astra for Law leads in legal intelligence
A new item from OpenAI. Follow the source link to read the complete publication.
Introducing OpenAI for Chile
A new item from OpenAI. Follow the source link to read the complete publication.
More researchers get access to ChatGPT
OpenAI expands its academic programme, with a new application round ahead.
Strengthening the fight against online child sexual abuse with new tools and support
A new item from OpenAI. Follow the source link to read the complete publication.
Science at the Frontier: How scientists use ChatGPT
A new item from OpenAI. Follow the source link to read the complete publication.
How Revolut raises ambition without lowering the bar
A new item from OpenAI. Follow the source link to read the complete publication.
LVMH’s Media Division announces strategic partnership with OpenAI
A new item from OpenAI. Follow the source link to read the complete publication.
Our first phase of embedded assessments
A new item from OpenAI. Follow the source link to read the complete publication.
Advancing computer-use agents for healthcare teams
A new item from OpenAI. Follow the source link to read the complete publication.
More ways to move care forward with ChatGPT for Healthcare
A new item from OpenAI. Follow the source link to read the complete publication.
The King’s Trust and OpenAI launch £1m partnership
A new item from OpenAI. Follow the source link to read the complete publication.
At Holiday Extras, dots get to work before anyone asks
A new item from OpenAI. Follow the source link to read the complete publication.
New Ask Maps features to help with holiday travel and prep
A new item from Google AI. Follow the source link to read the complete publication.
Hugging Face and OpenRouter join forces to make open-weight models easy to run
A new item from Hugging Face. Follow the source link to read the complete publication.
A step toward defining the Standard Model on a lattice
A new item from OpenAI. Follow the source link to read the complete publication.
Our new blog for science
A new item from OpenAI. Follow the source link to read the complete publication.
ChatGPT Wallet adds a payment step to agent workflows
The launch separates shopping assistance from a user-approved purchase.
Shield helps technicians hand off legacy portal work with OpenAI
A new item from OpenAI. Follow the source link to read the complete publication.
Build a Robot Pick-and-Place Simulation with SimReady Assets
A new item from Hugging Face. Follow the source link to read the complete publication.
XYBench: Helping with the problem behind the question
A new item from Hugging Face. Follow the source link to read the complete publication.
llama.cpp v1.0.0 : A stronger foundation for the local AI ecosystem
A new item from Hugging Face. Follow the source link to read the complete publication.
Gemini 4 Hydrogen targets faster execution
Google positions Hydrogen for coding and everyday agent workflows.
How Amplitude ships product improvements in 1–2 days with OpenAI
A new item from OpenAI. Follow the source link to read the complete publication.
Gondola doubles revenue as its team builds more with Codex
A new item from OpenAI. Follow the source link to read the complete publication.
Holiday Extras is rethinking what it can build with Codex
A new item from OpenAI. Follow the source link to read the complete publication.
Supporting research on AI and children’s safety
A new item from OpenAI. Follow the source link to read the complete publication.
ChatGPT Ads increases its market presence in 44 countries
A new item from OpenAI. Follow the source link to read the complete publication.
Gilbert + Tobin rethinks the business of law with AI
A new item from OpenAI. Follow the source link to read the complete publication.
Google Search gets a major upgrade with Gemini 4 Hydrogen
A new item from Google AI. Follow the source link to read the complete publication.
LLMON: An LLM-Native Markup Language, now in Granite 4.1 3B and 8B
A new item from Hugging Face. Follow the source link to read the complete publication.
The Definitive Fine-Tuning Guide for 3D Generation Models
A new item from Hugging Face. Follow the source link to read the complete publication.
An AI API for choices, not chat replies
Gemini Decisions scores predefined options across multimodal inputs.
Try Tabletop: Customizable, collaborative spaces for groups to plan, play and do more together.
A new item from Google AI. Follow the source link to read the complete publication.
Gemini 4 Carbon opens up for longer, harder tasks
Google positions Carbon around sustained reasoning and extended verification.
AlphaProof Code puts AI-generated proofs to work
DeepMind reports formal verification of key properties in Signal protocol code.
Claude Fable and Mythos reach version 5.2
Anthropic updates its models for demanding technical work.
Mistral Studio brings agents, workflows and apps together
The enterprise platform adds tools for building and observing AI systems.
Images 2.6 adds motion and editing controls
OpenAI expands image generation with GIFs and more precise iteration.
Mistral releases Le Chonk model weights
Mistral Large 4 moves beyond its preview with downloadable weights.
Gemini 4 Carbon begins with trusted testers
Google announces a new model for sustained engineering and cyber defense.
Genie 4 adds sound to worlds you can steer
Google DeepMind previews interactive simulation with native audio.
Agent Skills moves toward independent stewardship
Anthropic announced a donation to the Agentic AI Foundation.
Mistral and NVIDIA plan shared tests for engineering AI
The collaboration combines simulation tools with benchmarks for technical agent work.
Voxtral 3 joins speech and tools in one voice workflow
Mistral pairs new speech models with a real-time agent SDK.
Timestral brings foundation models to forecasting
Mistral's compact model family predicts numerical series rather than chat replies.
Gemini 4 Argon moves into broad availability
Google opens its frontier model across developer and subscriber products.
Haiku 5.5 aims at the small tasks that add up
Anthropic expands its inexpensive model line with adjustable reasoning effort.
GPT-6.1 Sol brings a lower-cost option for complex work
OpenAI targets coding, documents and computer workflows with an updated Sol model.
Opus 5.5 puts price and task completion in the same frame
Anthropic's new release targets coding and professional workflows at lower cost.
Anthropic brings evaluators inside the development process
An Accenture partnership proposes deeper access to models and deployment decisions.
Mistral models enter Firefox's Smart Window beta
Mozilla's browsing assistant adds another route for open-model AI to reach users.
A Fortran migration puts numerical checks ahead of code volume
Mistral describes an agent-assisted modernization of a scientific simulator.
GPT-6 Astra links reasoning with computer work
OpenAI begins a staged rollout focused on complex professional tasks.
Claude's 5.1 release splits access by safeguards
Anthropic updates Fable and Mythos while reducing the price of cached context.
Omni 1.1 Flash adds finer control over generated video
Scene extension, endpoint frames and lightweight previews expand Google's creative API.
Mistral gives retrieval a second look at the evidence
Agentic Search lets models navigate documents instead of stopping at initial matches.
Gemini 3.7 Flash targets cheaper coding and agent work
Google updates its workhorse model three weeks after the preceding Flash release.
Mistral adds regional control to its inference service
European and US endpoints arrive alongside a preview of committed service levels.
ChatGPT's August update adds a choice of reasoning effort
OpenAI documents new everyday models and an updated Sol experience.
Shieldstral lets developers describe moderation policies in words
Mistral releases a compact open model for text and image safety checks.
OpenAI publishes ten AI-assisted mathematics results
New arguments and formal certificates invite scrutiny beyond a model benchmark.
Gemini Robotics ER 2 coordinates actions and multiple robots
A higher-level model plans tasks while lower-level systems execute movement.
Opus 5 targets stronger everyday autonomous work
Anthropic positions its upgraded Opus closer to Fable-level capability.
Google refreshes Flash for cheaper, faster agent work
Three models cover general work, low latency and security workflows.
OpenAI reports failures found during longer agent runs
Internal deployment revealed behavior missed by earlier evaluations.
GPT-5.6 splits into Sol, Terra and Luna
Three model tiers target different balances of capability and cost.
Robostral Navigate follows instructions through a single camera
An eight-billion-parameter model targets navigation from ordinary RGB images.
Leanstral 1.5 pairs generated proofs with formal checks
An open model targets practical proof engineering in Lean 4.
Sonnet 5 brings stronger agents to a lower price tier
Anthropic targets tool use, planning and everyday knowledge work.
OCR 4 makes document structure available to agents
Bounding boxes and confidence scores join extracted document text.
Claude Tag brings shared agent work into Slack
Teams can delegate tasks from channels with shared context and controls.
Anthropic launches Fable 5 with a restricted Mythos counterpart
Two configurations share a model while applying different safeguards.
Codex adds Sites and tools for more kinds of work
Role-specific plugins, annotations and internal sites broaden Codex's scope.
Cosmos 3 combines world simulation and robot actions
An open model family connects visual reasoning, generated worlds and actions.
Opus 4.8 adds more control over effort
Anthropic pairs its model upgrade with effort controls and dynamic workflows.
Mistral expands into models for physical simulation
Engineering models aim to predict physical behavior from design inputs.
Mistral sends Vibe coding agents to the cloud
Remote coding sessions arrive alongside an open-weight model and Work mode.
Gemini 3.5 starts with Flash and a focus on action
Google launches a model family built around longer, tool-driven tasks.
Gemini Omni turns mixed inputs into editable video
A multimodal model connects generation with conversational video editing.
OpenAI brings reasoning to live voice
Three audio models target conversation, translation and streaming transcription.
Mistral Workflows adds recovery and approvals to AI processes
The Studio public preview focuses on durable execution and visible operational history.
GPT-5.5 targets tasks that cross tools and applications
OpenAI positions the release around sustained coding, research and office work.
Claude Opus 4.7 emphasizes verification on harder tasks
Anthropic updates software engineering and vision while introducing new cyber safeguards.
Muse Spark starts a new model family at Meta
Meta Superintelligence Labs introduces a multimodal reasoning model for its assistant.
Gemma 4 widens Google's open-model lineup
Four model sizes target local hardware, reasoning and tool-based workflows.
Lyria 3 Pro stretches AI music generation to three minutes
Google adds more control over song structure and expands access across its products.
Mistral Forge targets models trained on institutional knowledge
A new enterprise system supports customization across the training lifecycle.
Mistral Small 4 combines reasoning, vision and coding
An Apache-licensed release joins capabilities previously spread across model families.
Gemini Embedding 2 gives mixed media a shared search space
Google's preview connects text, images, audio, video and documents.
GPT-5.4 brings coding and office work into one model
OpenAI rolls the model into ChatGPT, Codex and the API.
Nano Banana 2 brings faster image creation and editing
Google's Flash Image update targets text rendering, visual consistency and iteration.
Gemini 3.1 Pro expands Google's reasoning model across products
A preview brings updated problem-solving to developers and consumer apps.
Sonnet 4.6 brings stronger computer use to Claude's everyday tier
Anthropic updates its default model for Free and Pro users.
Claude Opus 4.6 pushes toward longer coding and office tasks
Anthropic adds a million-token context beta and adaptive reasoning controls.
GPT-5.3-Codex combines coding with longer professional workflows
OpenAI expands the coding model's role in tool use and computer work.
Voxtral splits transcription into batch and live models
Mistral adds speaker labels and an open-weight real-time option.
AlphaGenome's published study maps DNA to regulatory signals
A Nature paper details predictions across long sequences and multiple biological measurements.
Prism brings AI into the scientific manuscript workspace
OpenAI launches a free LaTeX environment with GPT-5.2 and collaboration.
TranslateGemma puts specialized translation into smaller models
Google releases three model sizes evaluated across 55 languages.
Gemini starts connecting personal context across Google apps
An optional US beta links Gmail, Photos, YouTube and Search.
ChatGPT creates a separate space for health questions
The January announcement introduces a dedicated experience with additional data protections.
NVIDIA opens a reasoning toolkit for autonomous driving
Alpamayo pairs a driving model with simulation tools and research datasets.
No stories match this month and topic. Try another selection.