16:30
Skild AI uses NVIDIA physical AI for robot foundation model S1
Skild AI launched the S1 robot foundation model, which learns unseen long-horizon tasks from a single video demonstration. It uses NVIDIA physical AI technology.
1 official
The Wire
Launches
Linked and quoted from the labs, the press and the people using them. Times in UTC.
The Wire
The 17 labs we track and the biggest stories beyond them. The rest is in Everything.
Thursday 10 September
2 developments15:00
OpenAI launches Data agent in ChatGPT Work
OpenAI introduced a Data agent in ChatGPT Work. It lets users connect company data, uncover insights, and build interactive dashboards using natural language.
1 official
Wednesday 9 September
2 developments16:00
Google DeepMind AI used to recreate couple's past in short film
Filmmakers and Google DeepMind used AI to recreate a couple's unrecorded past in the short film "Love, Rendered." The film was announced on Google's official blog.
1 official
11:19
DeepSeek releases V4.1 Flash on Hugging Face
DeepSeek published the DeepSeek-V4.1-Flash model on Hugging Face under an MIT license. The model supports image-text-to-text and text generation.
1 official3 coverage2 community
DeepSeekLaunch · 9 Sep
DeepSeek-V4.1-Flash
New open-weight model
$0.10 / $0.50per M tokens, in / out
1Mcontext
Openweights
Tuesday 8 September
3 developments19:00
Meta launches Muse personal AI agent
Meta announced Muse, described as a secure and private personal AI agent that proactively helps people meet their goals and suggests ideas. The announcement appeared on Meta's official newsroom.
1 official4 coverage
18:37
OpenAI introduces ChatGPT Images 2.5
OpenAI announced ChatGPT Images 2.5, a new version of its image generation feature in ChatGPT. The announcement links to an OpenAI page introducing the update.
2 coverage1 community
15:11
Cohere releases tiny-aya-base-32K multilingual model
CohereLabs published tiny-aya-base-32K on Hugging Face, a multilingual long-context text-generation model. It is tagged aya, multilingual, long-context and tiny-aya.
1 official
CohereLaunch · 8 Sep
tiny-aya-base-32K
New open-weight model
32Kcontext
Openweights
Friday 4 September
1 development08:54
Tencent releases EVIE-4.5B vision language model on Hugging Face
Tencent published the EVIE-4.5B vision-language model on Hugging Face. The model uses ColBERT late-interaction and multi-vector retrieval for visual document retrieval.
2 official
TencentLaunch · 4 Sep
EVIE-8B
New open-weight model
Openweights
Thursday 3 September
2 developments18:18
OpenAI launches ChatGPT for Financial Services with GPT-6 Astra
OpenAI introduced ChatGPT for Financial Services, which combines built-in financial data with GPT-6 Astra. It is aimed at research, modeling, and client-ready materials.
1 official10 coverage2 community
OpenAILaunch · 3 Sep
GPT-6 Astra
New model generation from a lab
$10.00 / $50.00per M tokens, in / out
1.05Mcontext
15:00
Google DeepMind launches WeatherNext 3 weather AI model
Google DeepMind introduced WeatherNext 3, described as its most advanced and accurate global weather AI model. The announcement comes from the official DeepMind site.
1 official2 coverage
Wednesday 2 September
5 developments19:35
Meta releases Muse Spark 1.3 AI model
Meta's Muse Spark 1.3, a multimodal reasoning model for agentic and coding workflows, is listed on OpenRouter. It supports a 1,048,576 token context window and is priced at $1.25 per million input tokens and $4.25 per million output tokens.
5 coverage1 community
MetaLaunch · 2 Sep
Muse Spark
New flagship model
$0.10 / $0.20per M tokens, in / out
1Mcontext
16:24
Google DeepMind launches proactive cyber defense for governments and enterprises
Google DeepMind announced a proactive cyber defense offering aimed at governments and enterprises. The announcement was published on its official site.
1 official
15:40
Google launches Fairwind Program for cyber defense
Google announced the Fairwind Program, a limited access program for governments and trusted partners to use its cyber defense tools. The program is tied to Gemini 3.
1 official
15:22
Cohere releases tiny-aya-l2-thinker model on Hugging Face
CohereLabs published the tiny-aya-l2-thinker model on Hugging Face. It is a multilingual reasoning text-generation model in the tiny-aya family.
2 official
CohereLaunch · 2 Sep
tiny-aya-en-thinker
New open-weight model
Openweights
15:14
Google launches Gemini 3.8 Flash and 3.8 Flash Cyber
Google introduced Gemini 3.8 Flash and Gemini 3.8 Flash Cyber. Both are Flash-tier models from the Gemini 3.8 line.
1 official4 coverage
Tuesday 1 September
3 developments17:08
Google introduces agentic video understanding with Gemini
Google announced agentic video understanding capabilities with Gemini. The feature is presented as a new capability for the Gemini model line.
1 official1 coverage
17:03
Google publishes gnm-v3 head model on Hugging Face
Google released gnm-v3, a 3D parametric human head model for head tracking and digital humans, on Hugging Face. The model is tagged for computer vision, mesh, and 3DMM use.
1 official
Google DeepMindLaunch · 1 Sep
gnm-v3
New open-weight model
Openweights
16:00
Google launches Pics image tool in Workspace
Google Pics, an image creation and editing tool, is now available in Google Workspace. It is built on Google's latest Nano Banana model.
1 official1 coverage