EQ-Bench 4

Emotional intelligence in multi-turn role-play, judged pairwise into an Elo-style rating.

Published by

EQ-Bench

Category

Human preference

Index weight

0.5

Models

27

Data as of

9 Sept 2026

MIT — EQ-Bench.

1. 1Claude Opus 51385
2. 2Claude Fable 51340
3. 3Kimi K31339
4. 4GPT-5.51315
5. 5Claude Opus 4.71311
6. 6Claude Opus 4.81281
7. 7GPT-5.41272
8. 8Muse Spark 1.11260
9. 9GPT-5.6 Sol1250
10. 10Claude Sonnet 51236
11. 11GPT-5.6 Terra1234
12. 12Claude Opus 4.61223
13. 13GLM-5.21222
14. 14MiMo-V2.5-Pro1208
15. 15Claude Sonnet 4.61207

SearchAll providersAlibabaAnthropicDeepSeekGoogleMetaMiniMaxMistral AIMoonshot AINVIDIAOpenAIxAIXiaomiZhipu AIOpen weights27 of 27

#	Model	Rating ↓	Run ↕	Index ↕	Published ↕
1	Claude Opus 5Anthropic	1385	–	70.0	–
2	Claude Fable 5Anthropic	1340	–	70.4	–
3	Kimi K3Moonshot AIopen	1339	kimi-k3	66.5	–
4	GPT-5.5OpenAI	1315	gpt-5.5	66.6	–
5	Claude Opus 4.7Anthropic	1311	–	65.5	–
6	Claude Opus 4.8Anthropic	1281	–	64.7	–
7	GPT-5.4OpenAI	1272	gpt-5.4	63.4	–
8	Muse Spark 1.1Meta	1260	–	64.9	–
9	GPT-5.6 SolOpenAI	1250	–	65.3	–
10	Claude Sonnet 5Anthropic	1236	–	61.2	–
11	GPT-5.6 TerraOpenAI	1234	–	62.3	–
12	Claude Opus 4.6Anthropic	1223	–	63.1	–
13	GLM-5.2Zhipu AIopen	1222	glm-5.2	57.4	–
14	XMiMo-V2.5-ProXiaomiopen	1208	–	59.0	–
15	Claude Sonnet 4.6Anthropic	1207	–	60.3	–
16	Kimi K2.6Moonshot AIopen	1202	kimi-k2.6	60.6	–
17	DeepSeek V4 ProDeepSeekopen	1166	–	59.7	–
18	GPT-5.6 LunaOpenAI	1156	–	56.6	–
19	MiniMax-M3MiniMaxopen	1150	minimax-m3	56.9	–
20	Gemini 3.1 ProGoogle	1142	–	63.9	–
21	Gemma-4-31B-ITNVIDIAopen	1120	–	55.4	–
22	Qwen3 7Alibaba	1110	–	58.9	–
23	Gemini 3.5 FlashGoogle	1087	–	60.7	–
24	Grok 4.3xAI	1075	grok-4.3	57.8	–
25	Claude Haiku 4.5Anthropic	1064	–	47.9	–
26	Qwen3.6 27BAlibabaopen	1026	–	53.2	–
27	Mistral Medium 3.5Mistral AI	993	–	50.8	–