OpenAI released GPT-5.6 on July 9 across ChatGPT, its Codex coding tool and its developer interface, following a 13-day government preview. Instead of a single model, the company shipped a three-tier family named Sol, Terra and Luna.
Sol is the flagship, built for the hardest reasoning, coding and science work, priced at 5 dollars per million input tokens and 30 dollars per million output tokens. Terra is the balanced everyday option at 2.50 and 15 dollars, while Luna is the fast, low-cost tier aimed at high-volume tasks at 1 and 6 dollars.
The release adds several new features, including a Max reasoning setting, an Ultra mode that coordinates four AI agents working in parallel, and Programmatic Tool Calling, which lets the model write JavaScript in a sealed, offline runtime to organise its tools. OpenAI says that last feature cut token usage by between 38 and 63.5 percent for some customers.
On performance, Sol led the Artificial Analysis Coding Agent Index with a score of 80, ahead of rival Claude Fable 5’s 77.2, and scored 52.7 percent on the demanding Agents’ Last Exam, against Fable 5’s 40.5 percent. On the Terminal-Bench 2.1 test it reached 88.8 percent, rising to 91.9 percent when the parallel-agent Ultra mode was switched on.
The lineup was not without gaps. On the SWE-Bench Pro software test, Sol’s 64.6 percent trailed Claude Mythos 5’s 80.3 percent by roughly 15 points, showing that no single model yet leads on everything.
Why it matters: splitting GPT-5.6 into cheaper and more powerful tiers means users can pick a version they can actually afford, which matters for cost-conscious businesses and developers across Nigeria and the rest of Africa.
Comments 0