Side-by-side comparison of pricing, 12 benchmarks, and generation speed.
| Metric | ERNIE 4.5 300B A47B | Qwen3 8B (Reasoning) |
|---|---|---|
| Input ($/M tokens) | $0.28 | $0.11 |
| Output ($/M tokens) | $1.1 | $1.15 |
Data from Artificial Analysis API — 12 benchmarks
Qwen3 8B (Reasoning) is cheaper overall. Its blended price (3:1 input/output ratio) is $0.37/M tokens vs $0.48/M for ERNIE 4.5 300B A47B.
ERNIE 4.5 300B A47B wins 11 out of 12 benchmarks compared to 1 for Qwen3 8B (Reasoning). See the detailed benchmark chart above for per-category results.
Qwen3 8B (Reasoning) generates tokens faster at 86 tok/s vs 24 tok/s. However, Qwen3 8B (Reasoning) has lower time-to-first-token (1.03s vs 1.63s).
Choose based on your priorities: Qwen3 8B (Reasoning) for lower cost, ERNIE 4.5 300B A47B for stronger benchmark performance, and Qwen3 8B (Reasoning) for faster generation. For latency-sensitive apps, check the TTFT comparison above.