Report a run

Before you shareReports are private until reviewed. Likely secrets and local paths are redacted. Do not include prompts, outputs, or private machine details.

File and setup

Exact listed file

Hardware type

GPU vendor

GPU model

GPU VRAM · GB

System RAM · GB

CPU model · optional

Apple chip

Unified memory · GB

CPU model

System RAM · GB

Operating system (optional)

Driver/backend version (optional)

Runtime and task

Runtime

Runtime version/build (optional)

Backend

GPU layers offloaded (optional)

CPU threads (optional)

Configured context · tokens (optional)

Task category

What did you try?

Success criteria (optional)

Outcome

Measurements (optional)

Enter observed values only. Failures need no measurements. Prompt-processing rate is not time to first token.

Actual input tokens

Actual output tokens

Run state

Time to first token · seconds

Generation speed · tokens/s

Generation time · seconds

Prompt processing rate · tokens/s

Peak memory · GB

Memory pool measured

Memory measurement method (optional)

Notes (optional)

Cancel