Side-by-side comparison of pricing, 12 benchmarks, and generation speed.
| Metric | Phi-4 | DeepSeek V3.2 (Reasoning) |
|---|---|---|
| Input ($/M tokens) | $0.125 | $0.3 |
| Output ($/M tokens) | $0.5 | $0.45 |
Data from Artificial Analysis API — 12 benchmarks
Phi-4 is cheaper overall. Its blended price (3:1 input/output ratio) is $0.22/M tokens vs $0.34/M for DeepSeek V3.2 (Reasoning).
DeepSeek V3.2 (Reasoning) wins 11 out of 12 benchmarks compared to 1 for Phi-4. See the detailed benchmark chart above for per-category results.
Phi-4 generates tokens faster at 38 tok/s vs 0 tok/s. However, DeepSeek V3.2 (Reasoning) has lower time-to-first-token (0.00s vs 0.49s).
Choose based on your priorities: Phi-4 for lower cost, DeepSeek V3.2 (Reasoning) for stronger benchmark performance, and Phi-4 for faster generation. For latency-sensitive apps, check the TTFT comparison above.