Intel60.9
Code77.4
Agents57.8
OpenAI’s flagship. Leads ARC-AGI-2 and Terminal-Bench. Token-efficient per task even though list price looks high.
- Price /1M
- $5 / $30
- Context
- 1.05M
- Speed
- 74 t/s
- Status
- Current
Abstract reasoningCodex-style agents
San Francisco
GPT-5.6 is a three-tier stack (Sol / Terra / Luna). Sol leads abstract reasoning; Luna is the cheap high-volume option.
openai.comOpenAI’s flagship. Leads ARC-AGI-2 and Terminal-Bench. Token-efficient per task even though list price looks high.
The middle GPT-5.6. Keeps most of Sol’s coding, drops some agentic depth, much faster and cheaper.
Still excellent, now priced like Sol while scoring like Terra. Migrate unless a contract pins it.
High MMLU, mid coding. Useful as a history point. Not a 2026-Q3 default.
The punchline of the 5.6 stack: near-Terra coding at Luna prices. Default OpenAI pick for volume.
OpenAI’s open-weight line. Fine to self-host; do not confuse it with the 5.6 API stack.