OpenAI has launched GPT-5.6 models after the U.S. Department of Commerce approved public release, ending a restriction imposed since late June. The models outperform competitors on coding benchmarks; Sol Ultra hit 91.9 percent on TerminalBench 2.1, while pricing at $5/$30 per million tokens undercuts Anthropic's comparable offering. The approval, enabled by testing from the Center for AI Standards and Innovation, represents one regulatory path for releasing advanced models, though formal standards still lack binding force.
Summaries like this, in your inbox every morning.
Sign up free →What happened
OpenAI released GPT-5.6 models on Thursday following approval from the U.S. Department of Commerce. The models had been restricted to select partners under government pressure since their unveiling in late June, but the Center for AI Standards and Innovation's additional testing cleared the way for public launch.
Why it matters
OpenAI criticized the delay, arguing it kept the best tools away from developers and companies. The approval signals a path forward for releasing advanced models, though binding standards for such releases—as called for in Trump's latest AI executive order—still do not exist.
What to watch
GPT-5.6 Sol Ultra scored 91.9 percent on TerminalBench 2.1 coding benchmark, ahead of Anthropic's Claude Mythos 5 at 88.0 percent. Sol costs $5/$30 per million input/output tokens, compared with nearly double that for Anthropic's Fable 5.
OpenAI's GPT-5.6 launch marks the first major public release after a regulatory hold imposed by the U.S. government. The delay—spanning from late June through the approval announcement—reflected government pressure to evaluate advanced models before wider distribution. The Center for AI Standards and Innovation's testing role suggests a potential approval framework for future releases, though the body notes that binding standards still do not exist despite Trump's recent executive order calling for them.
The release carries practical implications for pricing and capability in the generative AI market. OpenAI's Sol Ultra's 91.9 percent score on TerminalBench 2.1 establishes clear performance leadership over Anthropic's Claude Mythos 5 (88.0 percent), while its $5/$30 per million token pricing undercuts Anthropic's Fable 5 at nearly double that rate. On cybersecurity tasks, Sol matched Mythos 5 while using only a third of the tokens—a meaningful efficiency difference for cost-sensitive deployments. These metrics position the release as a competitive stake-raiser in the market for advanced reasoning and coding models.
No comments yet. Be the first to share your thoughts!
Log in to join the discussion
Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.
Get Started FreeFree · takes 30 seconds · unsubscribe anytime
1 minute a day. The AI essentials.
200+ sources · Email / LINE / Slack
