Four models landed this week: Qwen 3.8 Max, Gemini 3.7 Flash, Grok 4.6, GLM 5.3.
Nobody should have to bet on which lab is ahead this month. That's the whole point of putting them behind one endpoint. Also: all our AI products are now under AI Services.
Fun Spaceship Scene (no audio) - I prompted a tiny LLM local PC model Qwen 3.8-27b to create this on a modest 6GB RTX 3050.
Open source LLMs are on fire right now.
GLM 5.3 especially: three times fewer parameters than Kimi K3, and it still comes out ahead. Could not resist seeing what that does in 3D. Ran it through Blender, Unity and Godot, and took Qwen 3.8 Max along because that one is a curious
First look at Qwen 3.8-Flash-Next on 2 DGX Sparks with sgl_project RadixArk’s NVFP4 in OMP
You have reached the end of the archive
All of qwen38