AISlashdotabout 4 hours ago
OpenAI Releases 750-task life science benchmark
2 min read
OpenAI released a 750-task benchmark to evaluate AI performance in real life science research. Its top model GPT-Rosalind passed only 36.1% of tasks, failing 63.9%. Results show AI struggles with complex experimental design when inputs are in natural language, not LaTeX.
Level
Hype check
Tap to vote and see what everyone thinks.
In this storyOpenAI
