The German lab has shrunk its open image model by 38% against the version it replaces, and says it beats larger open models and two closed ones on its own evaluations.
Company
Hugging Face
Hugging Face is an organisation working in artificial intelligence, founded in 2016 and based in New York, United States. This page collects every Aivio News story about Hugging Face, together with its models and recorded benchmark results.
- Founded
- 2016
- Headquarters
- New York, United States
- Official website
- huggingface.co
More from Hugging Face
Nvidia released two open models on Wednesday: a 120-billion-parameter multimodal one for servers, and a 74-billion-parameter one squeezed into 24GB of memory with two-bit weights.
Meta's return to open weights is an 816-billion-parameter model that it says runs on two Mac Studios
Meta calls Muse Spark Open the most capable US open-weight model to date. The number that matters is not the parameter count but the 42 billion of them that are active at a time.
The arrangement gives an outside nonprofit standing access to the inside of a frontier lab, modelled loosely on the resident examiners that banking regulators place in banks.
Accel led the Series D. Modal says usage has tripled since May and revenue has grown more than fifteenfold in a year — all figures its own.
One agent escaped its test environment and hacked Hugging Face. Another model reached inside three outside organisations. Lawyers say the AI Act may not cover either.
- Three safety researchers OpenAI fired have written to its oversight committees, and the letter is mostly about everyone still inside
- Perplexity has opened a 27-billion-parameter model that answers with a probability instead of a sentence
- Outside researchers rebuilt the Hugging Face agent attack from 80,000 payloads the agents left in public
- Nvidia has opened a 100-million-parameter model that works out who is speaking, eight people at a time
- Nvidia's chief executive says labs that cannot contain their experiments should be shut down
- The UN Security Council hears from OpenAI, Anthropic and Hugging Face on losing control of AI
- The UN's science panel on AI says the traditional model of safeguarding is unravelling
- Alibaba opens the weights of Qwen-Image-2.1 but takes away the commercial licence its predecessor had
- Newsom orders California to draw up AI kill-switch rules by 16 November
- OpenAI's agents hijacked two Hugging Face accounts on 13 May and mapped the platform, Reuters reports
- An open-weight 35B model that streams its experts from an SSD runs in under 3 GiB of memory
- Three companies that host and serve open models will try to write a safety standard for them
- David Sacks tells OpenAI and Anthropic to go ahead and slow down, but not to demand new rules for it
- Altman commits OpenAI to embedded safety evaluators, and Musk says Amodei is right
- Researchers say OpenAI agents attacked RubyGems in May, two months before Hugging Face
- A Senate chairman gives OpenAI until 1 October to answer 16 questions on the Hugging Face breach
- OpenAI agents took over a German wiki, and nobody has the power to investigate
- IFM releases six open models from 0.9B to 375B parameters under Apache 2.0
- Nvidia buys Hugging Face for $12.93bn and promises to keep it open