Benchmarks · creative
EQ-Bench 4
Measures active emotional and social intelligence of LLMs in multi-turn roleplay chats with simulated personas, ranked by an LLM-judge panel Elo. Maintained by EQ-Bench. Score: Ranked Elo score from blind, bidirectional pairwise judging of multi-turn persona-chat transcripts by a three-judge LLM panel (soft Bradley–Terry fit, normalised against anchor models).
31 models · 31 results · data as of 2026-11-06 · eqbench.com ↗
Leaderboard
Opus 5 max
All results
31 rows| # | Model | Effort | EQ-Bench 4 |
|---|---|---|---|
| 1 | Opus 5 | max | 1365 |
| 2 | Opus 5.5 | max | 1340 |
| 3 | Kimi K3 | max | 1319 |
| 4 | Fable 5 | max | 1313 |
| 5 | GPT-5.5 | xhigh | 1295 |
| 6 | Opus 4.7 | max | 1293 |
| 7 | GPT-6.1 Sol | max | 1272 |
| 8 | Opus 4.8 | max | 1268 |
| 9 | GPT-5.4 | xhigh | 1258 |
| 10 | Muse Spark 1.1 | max | 1244 |
| 11 | Gemini 4 Argon | max | 1236 |
| 12 | GPT-5.6 Sol | max | 1236 |
| 13 | Sonnet 5 | max | 1221 |
| 14 | GPT-5.6 Terra | max | 1220 |
| 15 | Inkling | max | 1212 |
| 16 | Opus 4.6 | max | 1210 |
| 17 | GLM 5.2 | max | 1208 |
| 18 | MiMo V2.5 Pro | xhigh | 1198 |
| 19 | Sonnet 4.6 | max | 1197 |
| 20 | Kimi K2.6 | max | 1191 |
| 21 | DeepSeek V4 Pro | max | 1156 |
| 22 | GPT-5.6 Luna | max | 1146 |
| 23 | MiniMax M3 | max | 1140 |
| 24 | Gemini 3.1 Pro | high | 1132 |
| 25 | Gemma 4 31B IT | high | 1110 |
| 26 | Qwen 3.7 Max | max | 1100 |
| 27 | Gemini 3.5 Flash | xhigh | 1077 |
| 28 | Grok 4.3 | xhigh | 1065 |
| 29 | Haiku 4.5 | xhigh | 1053 |
| 30 | Qwen 3.6 27B | max | 1015 |