OpenAI: GPT-5.6 Finally Launched After Government Blockade
Le brief IA que les pros lisent chaque soir
Les 7 actus IA du jour, décryptées en 5 min. Gratuit.
Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.
Choisis ton rythme
Gratuit · Pas de spam · Désabonnement en 1 clic
OpenAI: GPT-5.6 Finally Launched After Government Hold-up
OpenAI's GPT-5.6 models will be launched this Thursday, following a delay imposed by the U.S. government. Unveiled at the end of June, they were initially reserved for selected partners under pressure from the U.S. government. The Department of Commerce approved the public launch after the Center for Standards and Innovation in AI conducted additional testing, according to Axios. OpenAI has openly criticized this hold-up, stating that it prevents the best tools from reaching developers and businesses. Binding standards for the launch of such models, like those called for by the latest executive order on AI from Trump, still do not exist.
Model Performance
The GPT-5.6 Sol Ultra took first place in the coding benchmark TerminalBench 2.1 with a score of 91.9%. The Claude Mythos 5 achieved 88.0%, while Google's Gemini 3.1 Pro Preview finished with 70.7%.
OpenAI claims that Sol outperforms Anthropic's Claude Mythos 5 across several benchmarks. On TerminalBench 2.1, Sol scored 88.8%, Sol Ultra reached 91.9%, and Mythos 5 settled at 88%. In cybersecurity tasks, Sol matched Mythos 5 while using only one-third of the tokens. The cost of Sol is $5/$30 per million input/output tokens. In contrast, Anthropic's Fable 5 costs nearly double, at $10/$50, and likely consumes more tokens as well.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.