Side-by-side comparison of pricing, 12 benchmarks, and generation speed.
| Metric | DeepSeek V3.2 (Non-reasoning) | Hy3-preview (Reasoning) |
|---|---|---|
| Input ($/M tokens) | $0.28 | $0.123 |
| Output ($/M tokens) | $0.42 | $0.43 |
Data from Artificial Analysis API — 12 benchmarks
Hy3-preview (Reasoning) is cheaper overall. Its blended price (3:1 input/output ratio) is $0.20/M tokens vs $0.32/M for DeepSeek V3.2 (Non-reasoning).
Hy3-preview (Reasoning) wins 6 out of 12 benchmarks compared to 4 for DeepSeek V3.2 (Non-reasoning). See the detailed benchmark chart above for per-category results.
Hy3-preview (Reasoning) generates tokens faster at 204 tok/s vs 0 tok/s. DeepSeek V3.2 (Non-reasoning) also has lower time-to-first-token (0.00s vs 1.77s).
Choose based on your priorities: Hy3-preview (Reasoning) for lower cost, Hy3-preview (Reasoning) for stronger benchmark performance, and Hy3-preview (Reasoning) for faster generation. For latency-sensitive apps, check the TTFT comparison above.