Side-by-side comparison of pricing, 12 benchmarks, and generation speed.
| Metric | Qwen3 8B (Non-reasoning) | Llama 3.2 Instruct 90B (Vision) |
|---|---|---|
| Input ($/M tokens) | $0.18 | $0.72 |
| Output ($/M tokens) | $0.7 | $0.72 |
Data from Artificial Analysis API — 12 benchmarks
Qwen3 8B (Non-reasoning) is cheaper overall. Its blended price (3:1 input/output ratio) is $0.31/M tokens vs $0.72/M for Llama 3.2 Instruct 90B (Vision).
Qwen3 8B (Non-reasoning) wins 7 out of 12 benchmarks compared to 5 for Llama 3.2 Instruct 90B (Vision). See the detailed benchmark chart above for per-category results.
Qwen3 8B (Non-reasoning) generates tokens faster at 81 tok/s vs 42 tok/s. However, Llama 3.2 Instruct 90B (Vision) has lower time-to-first-token (0.37s vs 0.95s).
Choose based on your priorities: Qwen3 8B (Non-reasoning) for lower cost, Qwen3 8B (Non-reasoning) for stronger benchmark performance, and Qwen3 8B (Non-reasoning) for faster generation. For latency-sensitive apps, check the TTFT comparison above.