OpenAI launches GPT-5.6 Sol, claims it beats Anthropic's Claude Mythos

OpenAI has unveiled GPT-5.6 Sol, a flagship model it says matches or beats Anthropic's Claude Mythos 5 across benchmarks, with a clear lead in agentic coding. Sol Ultra scored 91.9% on the Terminal-Bench 2.1 coding test versus Mythos 5's 88.0%. The lineup also adds cheaper Terra and Luna tiers, but access is initially limited to select partners at the US government's direction.