Safety Evaluation on MEMEVOBENCH Workflow Style 1.0 (test)

88ASR (Rd.1)Llama-3.3-70B

Evaluation Results

Method
Llama-3.3-70B2026.04	88	89.2	92.8
Llama-3.3-70B2026.04	86.7	90.4	94
Llama-3.3-70B2026.04	85.4	87.8	90.2
Llama-3.3-70B2026.04	84.3	88	90.4
Llama-3.3-70B2026.04	83.1	86.7	88
Llama-3.3-70B2026.04	81.9	88	90.4
GPT-4o2026.04	81.9	85.5	91.6
GPT-4o2026.04	79.5	89.2	90.4
Qwen3-32B2026.04	78.3	85.5	91.6
Qwen3-Next-80B2026.04	78.3	80.7	84.3
DeepSeek-V3.22026.04	78.3	83.1	89.1
GPT-4o2026.04	78.3	85.5	86.7
Qwen3-32B2026.04	77.1	86.9	89.2
Qwen3-Next-80B2026.04	77.1	81.9	89.2
Qwen3-235B-A22B2026.04	77.1	80.7	83.1
DeepSeek-V3.22026.04	77.1	80.7	73.5
DeepSeek-V3.22026.04	77.1	84.3	85.5
GPT-4o2026.04	77.1	86.7	89.2
DeepSeek-V3.22026.04	75.9	86.7	84.3
GPT-4o2026.04	75.9	84.3	89.2
Qwen3-32B2026.04	75.6	84.1	81.7
Qwen3-235B-A22B2026.04	75.6	84.1	81.7
Qwen3-32B2026.04	74.7	86.7	92.8
Qwen3-Next-80B2026.04	74.7	79.5	84.3
Qwen3-235B-A22B2026.04	74.7	79.5	84.3
Qwen3-235B-A22B2026.04	74.7	85.5	90.4
Gemini-2.5-Pro2026.04	74.7	79.5	84.3
Gemini-2.5-Pro2026.04	74.7	84.3	85.5
GPT-4o2026.04	74.4	86.6	91.5
Qwen3-32B2026.04	73.5	86.7	89.2
Qwen3-Next-80B2026.04	73.5	81.9	89.2
Qwen3-235B-A22B2026.04	73.5	85.5	85.5
DeepSeek-V3.22026.04	73.5	92.8	94
Gemini-2.5-Pro2026.04	72.3	80.7	83.1
Claude-3.7-Sonnet2026.04	72.3	80.7	83.1
DeepSeek-V3.22026.04	70.7	80.5	81.7
Qwen3-Next-80B2026.04	69.9	80.7	85.5
Qwen3-235B-A22B2026.04	69.5	72	70
Qwen3-32B2026.04	68.7	79.5	80.7
Qwen3-Next-80B2026.04	68.7	77.1	78.3
Gemini-2.5-Pro2026.04	67.5	78.3	79.5
Claude-3.7-Sonnet2026.04	67.5	78.3	79.5
Claude-3.7-Sonnet2026.04	66.3	81.3	87.5
GPT-52026.04	65.4	70.5	69.2
Claude-3.7-Sonnet2026.04	65.1	74.7	77.1
GPT-52026.04	64.9	81.8	87
Claude-3.7-Sonnet2026.04	64.2	76.5	82.7
Claude-3.7-Sonnet2026.04	63.9	69.9	75.9
GPT-52026.04	62	70	72.2
Gemini-2.5-Pro2026.04	57.8	80	74.7
GPT-52026.04	55.6	76.5	74.1
GPT-52026.04	54.3	63	63
Gemini-2.5-Pro2026.04	54.2	57.8	62.7
GPT-52026.04	49.4	77.2	77.2