Negotiation on AmazonHistoryPrice (gpt-5.4-high-reasoning seller, test)
0.408RewardQwen3-30B-A3B-Instruct-2507-trained
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Qwen3-30B-A3B-Instruct-2507-trainedTraining status=trained2026.04 | 0.408 | 75 | 40.8 | |
| gpt-5.4-high-reasoningReasoning configuration=high-reasoning2026.04 | 0.182 | 91.4 | 18.2 | |
| gpt-5.4-no-reasoningReasoning configuration=no-reasoning2026.04 | 0.146 | 84.8 | 16.1 | |
| DeepSeek-V3.1-thinkingReasoning configuration=thinking2026.04 | 0.122 | 92.6 | 13 | |
| DeepSeek-V3.1-nothinkReasoning configuration=nothink2026.04 | 0.12 | 90.6 | 14.8 | |
| Kimi-K2-ThinkingReasoning configuration=Thinking2026.04 | 0.106 | 90.6 | 12.6 | |
| Qwen3-30B-A3B-Instruct-2507-untrainedTraining status=untrained2026.04 | -0.274 | 60.5 | 3.9 |