Benchmarks/GPT-5.6 Sol
OpenAI / Proprietary

GPT-5.6 Sol

A high-intelligence model that leads the cited Coding Agent Index snapshot in OpenAI’s Codex harness.

INTELLIGENCE59index
CODING80%
AGENTICNOT MEASURED
SPEED60tok/s
FIT / INTERPRETATION

Where the evidence points.

  • Coding agents
  • Deep reasoning
  • Large-context analysis
TECHNICAL CONTEXT

Score needs conditions.

Context
1M
Input / 1M
$5
Output / 1M
$30
Access
Proprietary
EVIDENCE TRAIL

Original measurements.

GPT-5.6 benchmark reportOPEN SOURCE ↗Model measurementsOPEN SOURCE ↗