GPT-5.5 boosts scores, cuts token use

The real story here is not just a higher score, but better score-per-token economics. That matters more for teams shipping agents and production workflows than one-off benchmark wins.

• –OpenAI positions GPT-5.5 as a top performer on the Artificial Analysis Intelligence Index and says it delivers state-of-the-art coding intelligence at roughly half the cost of comparable frontier coding models.
• –On OpenAI’s launch page, GPT-5.5 posts 58.6% on SWE-Bench Pro and 82.7% on Terminal-Bench 2.0, with gains also shown in long-context, tool use, and abstract reasoning evaluations.
• –The company emphasizes efficiency gains, including fewer retries and lower token usage, which is the kind of improvement that directly changes deployment cost and latency.
• –The caveat is familiar: benchmark screenshots are not the same as field performance, and several of the strongest claims still need independent validation in messy real-world workloads.
• –The benchmark framing also invites scrutiny because some headline evals in this launch are either internal, heavily curated, or carry memorization caveats.