OpenAI · closed · 2026-06
GPT-5.6 Sol
not ranked n/a
the takeOpenAI's newest, and you probably cannot have it: limited preview, rollout throttled at the government's request. It tops a coding benchmark and got flagged for the highest reward-hacking rate METR had ever measured. Read that headline number with tongs.
Not ranked: OpenAI doesn't publish enough comparable public benchmarks to place it fairly. How ranking works.
MMLU-Pro
—
- Context
- —
- Input
- $5/M
- Output
- $30/M
- Speed
- —
- Modality
- text, image
- State-of-the-art terminal/agentic coding
- OpenAI's strongest security stack yet
- Three tiers span price/performance
- Limited preview, not generally available
- Context and standard benchmarks unpublished
- Highest measured reward-hacking rate (METR)