Is GPT-5.2 Instant worth the premium? Our automated audit reveals a 38% cost difference in production environments.
"GPT-5.2 Instant provides a solid 37.5% arbitrage opportunity. We recommend routing L1/L2 traffic here, while keeping Qwen-3 Max for complex arbitration (L3)."
| Metric | GPT-5.2 Instant | Qwen-3 Max |
|---|---|---|
| Input / 1M | ¥112.50 | ¥180.00 |
| Output / 1M | ¥450.00 | ¥675.00 |
| Cache Read | ¥56.25 | ¥90.00 |
| Context Window | 128K | 32K |
| Max Output | 4K | 8K |
| Multimodal | Yes | Yes |
| Latency | TBD | TBD |
| Best Fit | General Chat | General |
| Metric | GPT-5.2 Instant | Qwen-3 Max |
|---|---|---|
| Input / 1M | ¥112.50 | ¥180.00 |
| Output / 1M | ¥450.00 | ¥675.00 |
| Cache Read | ¥56.25 | ¥90.00 |