Paste prompts, pick models, fan out in parallel. Results stream live into the matrix below.
Add prompts + pick at least one model above, then click Run eval.