GPT-6-sol
gpt-6-sol
OpenAI · chat · api
Intelligence
Context
1.1M
Max output
128K
Input $/1M
$5.00
Output $/1M
$30.00
Modalities
text
Released
14 Sept 2026
Intelligence Index via Artificial Analysis · 0–100, higher is better
AI summary
● machine-written
OpenAI releases GPT-5.6 Sol, flagship reasoning model for complex coding and agentic tasks
GPT-5.6 Sol is OpenAI's top-tier model in the GPT-5.6 family, designed for complex reasoning, coding, and agentic workflows. It achieves 88.8% on Terminal-Bench 2.1 (base mode) and 91.9% in Ultra high-effort mode, outperforming Claude Mythos 5 and GPT-5.5. The model supports a 1.1M token context window with up to 128K tokens of output.
What's new
- Three-tier family structure (Sol, Terra, Luna) replacing single model with adjustable effort
- Sol Ultra mode achieves 91.9% on Terminal-Bench 2.1 agentic coding benchmark
- Base Sol scores 88.8% on Terminal-Bench 2.1, 0.8 points ahead of Claude Mythos 5
- 1.1M token context window with 128K token maximum output
- Announced June 26, 2026 as limited preview with broader access planned
Best for
Complex reasoning and multi-step agentic tasksCommand-line workflows and agent-based codingLong-horizon problem solving in biology and cybersecurity