Qwen3.8-Flash-Next Coder with 768 routed experts removed: an MLX runner and a GGUF for a patched llama.cpp, both with optional SSD streaming.
🤝 Open to Collab
Alex Schiltmans
Schiltmans
AI & ML interests
Always asking what it takes to make this work, and why it works.
Speculative decoding · distilled drafters · architecture research
Recent Activity
updated a model 4 days ago
Schiltmans/Qwen3.8-Flash-Next-Coder-cut240-GGUF updated a model 4 days ago
Schiltmans/Qwen3.8-Flash-Next-Coder-cut240 liked a model 13 days ago
Schiltmans/Qwen3.8-Flash-Next-Coder-cut240-GGUFOrganizations
models 5
Schiltmans/Qwen3.8-Flash-Next-Coder-cut240-GGUF
Text Generation • 113B • Updated • 1.46k • 2
Schiltmans/Qwen3.8-Flash-Next-Coder-cut240
Text Generation • 57B • Updated • 13.5k • 2
Schiltmans/Ternary-Bonsai-2-27B-DFlash2-ft5
2B • Updated • 687 • 2
Schiltmans/Ternary-Bonsai-2-27B-DFlash2-ft5-GGUF
2B • Updated • 1.2k • 4
Schiltmans/Ternary-Bonsai-2-27B-DFlash2-ft5-mlx-4bit
2B • Updated • 361 • 2