Capability matrix
- Compute and format Q4_K_M · pinned local artifact · Ryzen Zen 5.
- Local entry point Short decode comparison.
- Server Unverified for this artifact. Text decoder
- Harness No recorded coding task. A tool profile is declared. Rebuild the serving bundle and run a real client task.
When to use it
The 120B-A6B Q4_K_M trial used a specific local 119,150,722,112-byte GGUF. The public four-shard UD download differs. The reported 40-token prompt and 64 comparisons used 8,192 capacity. Reserve this for an artifact-matched trial.
Configuration and tested scope · Publisher model card
Hardware and tuning
No comparable x86/ARM prefill and decode benchmark pair is recorded here. Full-checkpoint training, LoRA and other PEFT methods: no verified CKE recipe for this exact checkpoint in the reviewed evidence. For a trainable starter, use the FP32 and adapter recipes.