Side-by-side comparison of pricing, 12 benchmarks, and generation speed.
| Metric | Qwen3 4B (Non-reasoning) | DeepSeek V3.2 Exp (Non-reasoning) |
|---|---|---|
| Input ($/M tokens) | $0.11 | $0.275 |
| Output ($/M tokens) | $0.42 | $0.415 |
Data from Artificial Analysis API — 12 benchmarks
Qwen3 4B (Non-reasoning) is cheaper overall. Its blended price (3:1 input/output ratio) is $0.19/M tokens vs $0.31/M for DeepSeek V3.2 Exp (Non-reasoning).
DeepSeek V3.2 Exp (Non-reasoning) wins 11 out of 12 benchmarks compared to 1 for Qwen3 4B (Non-reasoning). See the detailed benchmark chart above for per-category results.
Qwen3 4B (Non-reasoning) generates tokens faster at 104 tok/s vs 0 tok/s. However, DeepSeek V3.2 Exp (Non-reasoning) has lower time-to-first-token (0.00s vs 0.98s).
Choose based on your priorities: Qwen3 4B (Non-reasoning) for lower cost, DeepSeek V3.2 Exp (Non-reasoning) for stronger benchmark performance, and Qwen3 4B (Non-reasoning) for faster generation. For latency-sensitive apps, check the TTFT comparison above.