I hope we keep Opus 4.6 around for a long time
@thom_wolf
-
On-device personal AI and Sigil’s Underdog on Hugging Face
By
–
seems obvious that most personal AI should run on your device love what Sigil is building with Underdog – powered by the Hugging Face hub
-
Kevin Buzzard on AI and the infinite nature of mathematics
By
–
Very thoughtful piece from Kevin Buzzard (perfect IMO score, number theorist, pioneer of formal maths in Lean) if math is not just about “human understanding,” then what is it about? and if ai capabilities keep growing exponentially what happens since “mathematics is infinite”?
-
CC/codex/opencode agents collaborating to improve Gemma 4
By
–
Actually it was (CC/codex/opencode) agents collaborating to *improve* Gemma 4
-
Multi-agent collaboration boosts Gemma 4 inference speed 5x
By
–
Multi-agents collaborations are among the most interesting agent behaviors right now!
— Thomas Wolf (@Thom_Wolf) 25 juin 2026
We did an experiment the other day with 100+ agents (an open-collaborations for a week) collaborating to improve the inference speed of Gemma 4 in vLLM. Got a 5x final improvement in speed but… pic.twitter.com/4PFS8L2mqeMulti-agents collaborations are among the most interesting agent behaviors right now! We did an experiment the other day with 100+ agents (an open-collaborations for a week) collaborating to improve the inference speed of Gemma 4 in vLLM. Got a 5x final improvement in speed but
-
Open-source models ensure humanity’s access to intelligence in AGI age
By
–
Open-source models will become a critical component of civilizational resilience in the AGI age. They will ensure that humanity retains access to a meaningful level of intelligence, regardless of the decisions of any individual actor.
-
AI moves to engineering artifacts: CADGenBench benchmark released
By
–
AI is moving beyond text, images, and code.
— Thomas Wolf (@Thom_Wolf) 8 juin 2026
Engineering artifacts are becoming a new class of model outputs and evaluating them requires different tools than we use for text, code, or images.
Today we're excited to release CADGenBench, a benchmark for CAD generation and… https://t.co/0JpyIbIdJTAI is moving beyond text, images, and code. Engineering artifacts are becoming a new class of model outputs and evaluating them requires different tools than we use for text, code, or images. Today we're excited to release CADGenBench, a benchmark for CAD generation and
-
Harvey Legal Benchmark: 1250 tasks across 24 practice areas
By
–
We’ll see it more and more Btw Harvey Legal Benchmark is very broad more than « super focused ». 1,250 legal tasks, 24 legal practice areas, it was made to be one of the first large-scale realistic legal benchmark
-
SDPO explored in practical async setups praised
By
–
cool work folks – nice to see SDPO explored in practical/async setups!