Mythos Catches Bugs By The Droves. GPT-5.4 Claims 83%. Trust But Verify.

Anthropic upgraded Fable and Mythos, with Mythos already catching software bugs in significant numbers within weeks of release. Meanwhile, OpenAI released GPT-5.4 barely three months after GPT-5.2, claiming it matches or outperforms human professionals 83% of the time in their own testing. Both moves target enterprise contracts, where models must handle complex professional tasks with minimal risk, delay, or cost.

This illustrates the enterprise trust cycle. AI companies need corporate contracts, which require demonstrated reliability on real work tasks, not just benchmark scores. The mechanism is agentic AI, models that act autonomously on multi-step professional work. The lesson: vendor claims mean nothing until third parties reproduce them. Wait for independent verification before betting your workflow on a number a company handed you.

Anthropic with Fable and Mythos upgrades, and OpenAI with GPT-5.4. Mythos is already demonstrating bug-catching capability in the field. OpenAI's 83% figure remains self-reported and unverified.