OpenAI's release of its first open-weights model since GPT-2 has been delayed due to a safety review, as stated by CEO Sam Altman. The best open model from the US this year is Meta's Llama 4, criticized for its lackluster reception. Other releases from companies like Microsoft, IBM, and Google have not been impressive compared to Meta’s large model. Meanwhile, China is making strides in AI development with public model releases, showcasing its strength in the field, particularly with the DeepSeek R1 model gaining popularity for its capabilities.
"While we trust the community will build great things with this model, once weights are out, they can't be pulled back. This is new for us, and we want to get it right."
Despite hundreds of billions of investment in GPUs, the best open model America has managed so far this year is Meta's Llama 4, which enjoyed a less than stellar reception and was marred with controversy.
China continues to excel in AI development with its open model releases, while US companies keep much of their advancements locked away.
In early 2025, DeepSeek became a household name following the release of its R1 model, a 671-billion-parameter LLM.
Collection
[
|
...
]