Two large language models were compared on a code-optimizing test: Claude Opus 4 achieved on average about a 3× speedup in May 2024, while the April Mythos Preview showed roughly a 52× improvement. The study found that a trained human needs 4–8 hours to reach a 4× speedup; the results may indicate an acceleration in model-assisted development.
Mythos Preview achieved a 52× speedup on a code-optimization test, Claude Opus 4 only ~3×
Two large language models were compared on a code-optimizing test: Claude Opus 4 achieved on average about a 3× speedup in May 2024, while the April Mythos Preview showed roughly a 52× improvement.
