Noometry

Writing & Preference benchmark

EQ-Bench 4 leaderboard

As of October 2026, Claude Opus 5 has the highest published EQ-Bench 4 score on Noometry at 1385, out of 28 models with results.

Last verified

About EQ-Bench 4

Emotional intelligence in multi-turn role-play conversations, such as building rapport and handling conflict, rated by pairwise LLM-judged comparisons.

Category
Writing & Preference
Introduced
2026
Format
Pairwise LLM-judged conversations
Unit
Arena rating (Bradley–Terry)
Official site
eqbench.com

Top 15 models

  1. Claude Opus 5 1385
  2. Claude Fable 5 1340
  3. Kimi K3 1339
  4. GPT-5.5 1315
  5. Claude Opus 4.7 1311
  6. Claude Opus 4.8 1281
  7. GPT-5.4 1272
  8. Muse Spark 1.1 1260
  9. GPT-5.6 Sol 1250
  10. Claude Sonnet 5 1236
  11. GPT-5.6 Terra 1234
  12. Inkling 1226
  13. Claude Opus 4.6 1223
  14. GLM-5.2 1222
  15. MiMo-V2.5-Pro 1208

Sponsored placements are available on pages like this one. Advertise on Noometry

All results

#ModelProviderRatingSettingSourceDate
1Claude Opus 5 Anthropic1385EQ-Bench
2Claude Fable 5 Anthropic1340EQ-Bench
3Kimi K3 Moonshot AI1339EQ-Bench
4GPT-5.5 OpenAI1315EQ-Bench
5Claude Opus 4.7 Anthropic1311EQ-Bench
6Claude Opus 4.8 Anthropic1281EQ-Bench
7GPT-5.4 OpenAI1272EQ-Bench
8Muse Spark 1.1 Meta1260EQ-Bench
9GPT-5.6 Sol OpenAI1250EQ-Bench
10Claude Sonnet 5 Anthropic1236EQ-Bench
11GPT-5.6 Terra OpenAI1234EQ-Bench
12Inkling Thinking Machines Lab1226EQ-Bench
13Claude Opus 4.6 Anthropic1223EQ-Bench
14GLM-5.2 Z.ai (Zhipu)1222EQ-Bench
15MiMo-V2.5-Pro Xiaomi1208EQ-Bench
16Claude Sonnet 4.6 Anthropic1207EQ-Bench
17Kimi K2.6 Moonshot AI1202EQ-Bench
18DeepSeek V4 Pro DeepSeek1166EQ-Bench
19GPT-5.6 Luna OpenAI1156EQ-Bench
20MiniMax-M3 MiniMax1150EQ-Bench
21Gemini 3.1 Pro Preview Google1142EQ-Bench
22Gemma 4 31B IT Google1120EQ-Bench
23Qwen3.7 Max Alibaba (Qwen)1110EQ-Bench
24Gemini 3.5 Flash Google1087EQ-Bench
25Grok 4.3 xAI1075EQ-Bench
26Claude Haiku 4.5 Anthropic1064EQ-Bench
27Qwen3.6 27B Alibaba (Qwen)1026EQ-Bench
28Mistral Medium 3.5 Mistral AI993EQ-Bench

Compare the leaders

Other writing & preference benchmarks

Frequently asked questions

What does EQ-Bench 4 measure?

Emotional intelligence in multi-turn role-play conversations, such as building rapport and handling conflict, rated by pairwise LLM-judged comparisons.

Which model has the highest EQ-Bench 4 score?

As of October 2026, Claude Opus 5 has the highest published EQ-Bench 4 score on Noometry at 1385, out of 28 models with results.

What is the best open-weight model on EQ-Bench 4?

Kimi K3 has the highest EQ-Bench 4 rating among open-weight models at 1339, ranking 3 of 28 overall.