Compliance evaluation on EU AI Act synthetic cases

89.16Accuracy

Gemini-2.5-Flash (ContextLens)

Updated 4d ago

Evaluation Results

Method
Gemini-2.5-Flash (ContextLens) 2026.04	89.16	88.5
GPT-4o-mini (ContextLens) 2026.04	88.11	87.53
GPT-4o (ContextLens) 2026.04	87.5	87.02
Llama-3.1-8B-Instruct (ContextLens) 2026.04	86.6	86.21
GPT-5 (ContextLens) 2026.04	86.33	89.05
DeepSeek-R1 (671B) (ContextLens) 2026.04	86.31	85.87
o3-mini (ContextLens) 2026.04	85.09	84.48
o3-mini 2026.04	84.83	84.36
ContextReasoner-7B-SFT 2026.04	84.33	83.73
ContextReasoner-7B-PPO 2026.04	84.33	83.65
DeepSeek-R1 (671B) 2026.04	84	83.3
GPT-5 2026.04	80.33	79.8
Gemini-2.5-Flash 2026.04	80	79.48
o3-mini 2026.04	77.75	75.14
GPT-4o 2026.04	75.33	73.39
GPT-5 2026.04	74.83	72.07
GPT-4o-mini 2026.04	74.33	73.62
GPT-4o-mini 2026.04	71.16	70.92
OpenThinker-7B 2026.04	70.5	69.72
Llama-3.1-8B-Instruct 2026.04	58.67	59.36
Qwen2.5-7B-Instruct (ContextLens) 2026.04	54.63	53.88
Qwen2.5-7B-Instruct 2026.04	46.33	41.89