Verdict
Built from the measurements below
- Accuracy is too close to call: GPT-5.6 Terra scored 75.0% and GPT-6 Astra 76.2% on the same 80 questions, a gap within what chance alone could produce.
- GPT-5.6 Terra is 2.1× cheaper per question as billed ($0.0063 vs $0.013), reasoning tokens included.
- GPT-5.6 Terra generates text 2.1× faster: 101 tokens per second against 49.
- GPT-5.6 Terra starts answering sooner: 1.1 s to the first token, against 1.3 s for GPT-6 Astra.
Which should you choose?
Choose GPT-5.6 Terra for the lower cost per correct answer ($0.0084 against $0.017), faster streaming and a quicker first token. Choose GPT-6 Astra for shorter outputs (about 205 tokens per question against 472). The widest subject gap is in medicine: GPT-6 Astra answered 6 of 11 correctly, GPT-5.6 Terra 4 of 11. With so few questions per subject, treat this as a lead to test on your own prompts, not a result.
Accuracy gap -1.2 points (z = -0.18). We call a winner only when z is at least 1.96, the usual 95% threshold. Speed, latency and price need a 10% margin.