This is a sad day for the AI community… I'm working on a benchmark we will soon release… and these are early results gathering pretty much all models out there, closed and opened. Red = closed source models
y axis = Elo score
x axis = release date
size = task cost … and
BIG TECH
-
Early benchmark results show closed source AI models leading
By
–
-
OpenAI and Broadcom unveil the Jalapeño chip for LLM
By
–
NEW OPENAI CHIP! OpenAI and Broadcom have just announced the fruit of their collaboration, Jalapeño, their new chip for LLM inference, designed in only 9 months! The chip adapts to the inference needs of OpenAI's roadmap, obtaining,
-
OpenAI and Broadcom Unveil Revolutionary AI Chip
By
–
OpenAI has announced its first AI chip, designed and produced in partnership with Broadcom. > New state-of-the-art in performance per watt.
> OpenAI models were used to accelerate its development.
> Will be deployed at gigawatt scale. -
OpenAI unveils Jalapeño, its custom AI chip for LLM inference
By
–
OpenAI just unveiled Jalapeño, its first custom AI chip built from the ground up for LLM inference. It's OpenAI going deeper into the full stack: chips, cores, memory, network, racks, scheduling, deployment, and product experience.
-
Microsoft’s MAI-Thinking-1 avoids third-party synthetic data and AI content
By
–
Most AI model releases focus on benchmarks. This one is more interesting because of what Microsoft says it did not use. With MAI-Thinking-1, Microsoft claims no third-party LLM-generated synthetic data during pre-training, active filtering of AI-generated content, and no hidden
-
Baidu’s Unlimited-OCR weights on Hugging Face
By
–
Weights → https://
huggingface.co/baidu/Unlimite
d-OCR
… -
Baidu’s Unlimited-OCR transcribes books in one pass, surpassing page-by-page models
By
–
BAIDU JUST DROPPED AN ABSOLUTE GAME-CHANGER FOR DOCUMENT AI
— Charly Wargnier (@DataChaz) 24 juin 2026
It’s called `Unlimited-OCR`, and it can literally transcribe an entire book in a single pass 🤯
Most vision models read a single page, forget the context, and eventually hit a wall where performance degrades and… pic.twitter.com/KUHrWFHYTWBAIDU JUST DROPPED AN ABSOLUTE GAME-CHANGER FOR DOCUMENT AI It’s called `Unlimited-OCR`, and it can literally transcribe an entire book in a single pass Most vision models read a single page, forget the context, and eventually hit a wall where performance degrades and
-
AI talent underpaid: each worth $150B as Google market cap fell 7%
By
–
AI talent is still highly underpaid: Google’s market cap fell by 7% when Noam Shazeer and John Jumper left, so on average they’re each worth $150B .
-
Street View Grounding Now Available for Google’s Offline Video Models
By
–
street view grounding now available for google’s offline video models. can’t wait till you can do “spatial RAG” to load in the right panoramas to reference – suddenly large scale real world locations become movie sets! https://t.co/AR3gLfj0p6
— Bilawal Sidhu (@bilawalsidhu) 23 juin 2026street view grounding now available for google’s offline video models. can’t wait till you can do “spatial RAG” to load in the right panoramas to reference – suddenly large scale real world locations become movie sets!
-
Anthropic client sues US government for restricting AI model access
By
–
Interesting: an Anthropic client has just sued the US government over its directive that forced Anthropic to cut off access to its AI models Fable 5 and Mythos 5.