2026
Qwen Ships Faster Qwen-Image Turbo
The accelerated 7B checkpoint that cuts generation down to eight denoising steps is downloadable now
Read updateThe accelerated 7B checkpoint that cuts generation down to eight denoising steps is downloadable now
Read updateA Google image model name surfaces on a third-party marketplace, with no Google page using it
Read updateMistral's largest model yet is callable today, but its weights stay locked until the end of the month
Read updateQwen's image model now has an official blog post, Hugging Face weights and third-party workflow support
Read updateAnthropic's safeguard-heavy twin of Fable 5.1 stays inside trusted access programs instead of general release
Read updateA Flash-preview build of MiniMax's coding and research model now has its own page on the vendor's site
Read update