Verdict
Built from the measurements below
- Accuracy is too close to call: GPT-6 Luna scored 72.5% and gpt-oss-120b 58.8% on the same 80 questions, a gap within what chance alone could produce.
- gpt-oss-120b is 2.0× cheaper per question as billed ($0.00013 vs $0.00025), reasoning tokens included.
- GPT-6 Luna generates text 1.7× faster: 137 tokens per second against 82.
- gpt-oss-120b starts answering sooner: 0.5 s to the first token, against 1.2 s for GPT-6 Luna.
- GPT-6 Luna accepts more context: 1.1M tokens against 131k.
Which should you choose?
Choose GPT-6 Luna for faster streaming and a 1.1M-token context window. Choose gpt-oss-120b for the lower cost per correct answer ($0.00021 against $0.00034), a quicker first token, open weights you can run yourself and more hosts to choose from on OpenRouter (23 against 3). The widest subject gap is in law: GPT-6 Luna answered 3 of 6 correctly, gpt-oss-120b 1 of 6. With so few questions per subject, treat this as a lead to test on your own prompts, not a result.
Accuracy gap +13.7 points (z = 1.83). We call a winner only when z is at least 1.96, the usual 95% threshold. Speed, latency and price need a 10% margin.