Capability matrix
- Compute and format W4A4 NVFP4 · Ryzen 9 9950X3D short bring-up.
- Local entry point Specialized text runner.
- Server Unverified for this artifact. Check for a service that accepts this artifact.
- Harness No recorded client task. Run one identified request, then a task with a checked result.
Choose the download
Publisher model card · CKE input files
Publisher task: image text to text. CKE example: W4A4 packed NVFP4. Use this artifact identifier:
hf://CohereLabs/command-a-plus-05-2026-w4a4.
Get started
Experimental 160 GB-class NVFP4 SafeTensors lane; use the exact evidence boundary below before quoting support Use the model-specific setup.
What has actually run?
Native packed NVFP4 runtime maps 675/675 required weights, passes leaf parity, and produces coherent short text. 126-token prompt plus 32 decode steps.
CPU evidence. Ryzen 9 9950X3D short bring-up; full-model and long-context parity open.
Performance. The cited model result does not give a comparable prefill/decode pair here. ARM throughput for this exact artifact is unverified here. Benchmark evidence · 32K to 128K results
Can I fine-tune it?
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.