Topic

#Ai Evaluation

3 articles on Ai Evaluation — news, releases, guides and analysis from the SourceFeed engine.

Frontier Coding Models Are Two Points Apart. Costs Aren't

SWE-rebench's decontaminated five-language leaderboard shows accuracy converging while price per solved issue diverges fivefold.

Anthropic’s 300K-conversation study turns model choice and locale into measurable behavior, not flavor.

The same PDF swings from 66 to 99 across runs, and the reason isn't a bug you can prompt away.