BenchLeader
OpenAIAuto-detected

GPT-5.6 Sol

Not yet ranked: too few independent results so far. Released 9 Jul 2026.

Blended price
Output speed
First answer
Context
How it scores by categoryDashed line = average model (50). One step of 15 = one standard deviation.
  1. Coding77
  2. Agents & tools67
  3. Maths70

Benchmark results

One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.

Coding

BenchmarkdefaultSource
WeirdML89.4%#7WeirdML

Agents & tools

BenchmarkdefaultSource
APEX-Agents40.0%#9Mercor

Maths

BenchmarkdefaultSource
FrontierMath Tier 480.5%#10Epoch AI Benchmarking Hub

See also

Data as of 9 Sept 2026. Compare with another model.