samuelcardillo/Qwen3-Coder-Next-Opus-4.6-Reasoning-Distilled

Fix architecture: MoE (~80B total, ~3B active) not dense

Upload full fine-tuned model (BF16 safetensors)

Fix context length (256K not 1M), add detailed test results

Update model card with benchmarks and training details

Add model card

initial commit