Verdict
Built from the measurements below
- Accuracy is too close to call: DeepSeek V4 Pro 0813 scored 71.2% and GPT-5.4 72.5% on the same 80 questions, a gap within what chance alone could produce.
- DeepSeek V4 Pro 0813 is 3.1× cheaper per question as billed ($0.0036 vs $0.011), reasoning tokens included.
- Output speed is similar (56 vs 55 tokens per second).
- DeepSeek V4 Pro 0813 starts answering sooner: 0.8 s to the first token, against 1.7 s for GPT-5.4.
Which should you choose?
Choose DeepSeek V4 Pro 0813 for the lower cost per correct answer ($0.0051 against $0.016), a quicker first token, open weights you can run yourself and more hosts to choose from on OpenRouter (23 against 3). Choose GPT-5.4 for shorter outputs (about 706 tokens per question against 1,318). The widest subject gap is in engineering: GPT-5.4 answered 10 of 15 correctly, DeepSeek V4 Pro 0813 8 of 15. 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.