's DeepSeek R1, Microsoft's Phi MoE, nvidia's Nemotron, LiquidAI's LFM2.5 and Qwen3.8 27B, all running on a laptop's Qualcomm Hexagon NPU.
None of them were built for it. 57 more across language, speech, vision and embeddings, all ready to drop into your app.
Gemini 3.8 Flash is sitting free in AI Studio right now.
Logged into every route this week. ~ MiniMax Agent: daily check-in drops 400 points, valid 30 days, no card. runs MiniMax-M3. ~ AI Studio: Gemini 3.8 Flash in the free tier. same model id devs pay
The above video was made with research by Grok feed into Gemini Notebook.
Here is the same process and initial research questions, using DeepSeek, the Chinese AI in place of Grok.
Holy shit this TCC comes out today
You have reached the end of the archive
All of deepseek