2/ Picking a model
Choose a popular local model, e.g. qwen-3.8-27b set your hardware on to see which quantizations run best on your machine now when you open a hf model card, it will suggest the best quantization level
1-Bit 177B or 4-Bit 27B?
The Dumb Big vs Smart Small test! Can a big 177B Qwen 3.8 Flash Next at IQ1_S quant beat Qwen 3.8 27B at IQ4_XS? Tested logic reasoning, code gen, and deterministic prompts. The result was 🔥
Run this prompt in your highest intelligent model.
I already ran it in Astra and Fable. Now running it in Qwen 3.8 27B uncensored by SpaceTimeViking What is the highest leverage thing humanity still doesn’t know that it could know in the next few decades? The question whose
GLM 5.3, GLM 5.3 Flash, and Qwen 3.8 Flash all for FREEEEEEEE
Full flagships at $0, each with a 1M context window. 150M tokens across the free lineup on signup. What you get: → GLM 5.3 Free → GLM 5.3 Flash Free → Qwen 3.8 Flash Free → 150M tokens to spend across all three
You have reached the end of the archive
All of qwen38