Image
Claude Opus 4.8 is here: Anthropic released a model that beats GPT-5.5 and for the first time can say "I'm not sure about that"
Anthropic has released Claude Opus 4.8 — a flagship model that outperforms GPT-5.5 across most benchmarks. But the biggest breakthrough isn't just raw performance: Opus 4.8 has learned to admit uncertainty, which is more essential for professional AI deployment than another point or two on a benchmark chart.