Verdict
Built from the measurements below
- Accuracy is too close to call: GPT-5.3-Codex scored 82.5% and GPT-6 Luna 72.5% on the same 80 questions, a gap within what chance alone could produce.
- GPT-6 Luna is 42× cheaper per question as billed ($0.00025 vs $0.01), reasoning tokens included.
- GPT-6 Luna generates text 1.4× faster: 137 tokens per second against 95.
- GPT-6 Luna starts answering sooner: 1.2 s to the first token, against 1.4 s for GPT-5.3-Codex.
- GPT-6 Luna accepts more context: 1.1M tokens against 400k.
Which should you choose?
Choose GPT-6 Luna for the lower cost per correct answer ($0.00034 against $0.013), faster streaming, a quicker first token, shorter outputs (about 431 tokens per question against 699) and a 1.1M-token context window.
Accuracy gap +10.0 points (z = 1.51). We call a winner only when z is at least 1.96, the usual 95% threshold. Speed, latency and price need a 10% margin.