Qwen3.8-2.4T-A95B and DeepSeek-V4-Pro-0813 (1.7T) also released, but unlike Gemini, Qwen and DeepSeek have also released the weights, this means that these models can be run locally, if you have the hardware.
Maybe one should support companies that release the weights of their models?
And Qwen3.8-27B release tomorrow 15h UTC+0, this is a model that many/most are able to run locally.
More news and discussion here: reddit.com/r/LocalLLaMA
Google has launched Gemini 3.7 Flash, promising major gains in coding, web development, and AI agents. The new model also comes with significantly lower introductory pricing, making advanced AI workflows cheaper to run.
English
Español
Français
Italiano
Nederlands
Polski
Português
Русский
Türkçe
Svenska
Chinese
Magyar