TokenPost topic

HumanEval

1 TokenPost story is tagged HumanEval. The newest appear first.

Latest HumanEval stories

• Modified DeepSeek Harness Uses 87.93% Fewer Tokens in Coding Test

  Its success rate also rose on the 98-task HumanEval benchmark. The token reduction reflects several changes to the agent setup.

  Nov 2, 2026, 4:23 AM EST · Technology