Multi-IF

Multi-IF - multi-turn instruction following benchmark evaluating complex multi-constraint instruction adherence.

Category

Instruction Following

Max Score

100.0

Score Type

percent

Active

Yes

Model Scores

Scores for Multi-IF

Model Score Date Verified
Qwen3.5-4B
88.80%
— Verified
granite-4.2-3B
94.93%
— Verified
Nemotron-3-Nano-4B
68.27%
— Verified
Gemma4-E4B
14.93%
— Verified
LFM2.5-8B-A1B
82.93%
— Verified
MiniCPM5-2B
84.00%
— Verified
LFM2.5-2.6B
97.33%
— Verified
Qwen3.5-2B
44.80%
— Verified
Gemma4-E2B
40.30
— Verified
Qwen3-Next-80B-A3B-Thinking
100.00%
10.09.2025 Verified
Qwen3-VL-30B-A3B
68.80%
04.10.2025 Verified