EIBench (Emotion Management) - DeepSeek-V4-Pro Simulator: leaderboard
EIBench (2026) is a simulator-based benchmark for interactive emotion management: an LLM plays the user in a multi-turn scenario and updates an emotion and relation state after every reply, and the final state is scored from -100 to +100 relative to scenario-specific start, success and failure anchors. This board is the scenario-weighted average over the 213 test scenarios with DeepSeek-V4-Pro as the user simulator instead of Qwen3-Max; higher is better.
Metric: Anchor-based outcome score x 100 (-100 to +100): the simulated user's final emotion and relation state mapped per axis to 0 at the start anchor, +1 at the success anchor and -1 at the failure anchor (clipped), scenario-weighted average over the four scene types; 213 held-out test scenarios, multi-turn dialogue with an LLM user simulator, DeepSeek-V4-Pro as the simulator; higher is better. Source: arxiv.org. Saturation forecast: Around August 2027. 15 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 4.6 | 24 | #161 |
| 2 | DeepSeek V4 Pro | 23.3 | #188 |
| 3 | GPT-5.4 | 22.7 | #123 |
| 4 | Kimi K2.6 | 21.9 | #178 |
| 5 | Gemini 3.1 Pro (Preview) | 21.5 | #94 |
| 6 | Qwen 3.6 Max Preview | 19.9 | #159 |
| 7 | GLM-5.1 | 19.4 | #179 |
| 8 | MiniMax-M2.5 | 17.6 | #582 |
| 9 | Seed 2.0 Pro | 17 | #211 |
| 10 | Gemini 3 Flash | 16.5 | #164 |
| 11 | Qwen 3 Max | 14.8 | #403 |
| 12 | Grok 4.20 | 13 | #297 |
| 13 | Grok 4.20 (Reasoning) | 10.6 | #297 (Grok 4.20) |
| 14 | Qwen 3 32B | 5.8 | #960 |
| 15 | Qwen 3 8B | -23.7 | #1830 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=eibench-emotion-management-deepseek-v4-pro-simulator · How It Works · Data refreshed daily, snapshot 2026-11-13.