10 LLM families (40 repos: W4A16, FP8, NVFP4, BF16) with 3:1 Gated DeltaNet + Quest Exact Memory: 1M context in ~4GB KV on vLLM.
-
defnoteatin/PseudoBrain-Qwen3.8-Flash-Next-NVFP4
Text Generation • 67B • Updated • 873 -
defnoteatin/PseudoBrain-Qwen3.8-Flash-Next-W4A16
Text Generation • 79B • Updated • 736 -
defnoteatin/PseudoBrain-Qwen3-Coder-Next-W4A16
Text Generation • Updated • 548 -
defnoteatin/PseudoBrain-Qwen3-Coder-Next-NVFP4
Text Generation • 46B • Updated • 568