Claude Shares Terminal Benchmark Scores Versus Fable 5

The official Claude account on X stated that the model sets a new standard across benchmarks. It reported 52.6 percent on Terminal-Bench-Science 0.1, more than double the score for Fable 5. The same post listed 55.8 percent on Terminal-Bench 4.0 compared with 42.0 percent for Fable 5. The account attached an image to the post. These figures come directly from the public reply issued by the verified account for the AI assistant built by Anthropic. No further details on the model version or testing conditions appear in the visible post.