Atlas

Benchmarks

← All benchmarks

EQ-Bench 3: Analytic

Chat & Writing · 2025-04-28

Within EQ-Bench 3's emotional-intelligence scenarios, this subscore rates the analytic dimension of model responses on a 0–10 scale. Higher values are better.

Top models (higher is better)

ModelScore
Kimi K2.69.5
Claude Fable 59.4
Kimi K2.59.4
Gemini 3 Pro Preview9.3
GPT-5.29.3
GPT-5.49.3
Gemma 4 26B A4B IT9.2
Opus 4.79.1
Horizon Alpha9.1
DeepSeek-V4-Pro9.0
Gemma 4 31B IT9.0
GPT-5.59.0
Kimi K2 Instruct9.0
Gemini 3.1 Pro Preview8.9
GLM-4.78.9
GLM-5.28.9
Sonnet 4.58.8
Sonnet 4.68.8
GPT-5.18.8
gpt-oss-120b8.8
Opus 4.68.7
Opus 4.88.7
DeepSeek-R18.7
Grok 4.1 Fast8.7
Hermes 4 405B8.6
o38.6
Gemini 2.5 Pro Preview 06-058.5
Gemma 4 12B IT8.5
GLM 4.7 Flash8.5
GLM-5.18.5
GPT-5 Chat (2025-08-07)8.5
Hivemind-32B-Preview8.5
Qwen3.5 397B A17B8.5
QwQ-32B8.5
GLM-58.4
Claude 3.5 Sonnet (Oct 2024)8.3
Sonnet 48.3
GPT-5.3 Instant8.3
Llama 3.1 Nemotron Ultra 253B V18.3
Mistral Small 3.2 24B Instruct 25068.3
Loading Atlas data…