Claude Mythos Leads 17 of 18 Benchmarks as Anthropic, Meta, and OpenAI Advance Frontier Models

Summary

Anthropic's Claude Mythos model has reportedly achieved leading performance on 17 of 18 benchmarks the company measured, establishing a new high-water mark for frontier model capability. Meanwhile, Meta's Muse Spark model has returned Meta to competitive standing among frontier AI labs, and OpenAI's internally codenamed 'Spud' model is reportedly nearing launch, signaling another wave of capability upgrades across the major model providers. This three-way advancement underscores the acceleratin