RecentsView All

Your evaluations

Sign in to run benchmarks and follow your results here.

Benchgen

Overview Leaderboard Evaluations Phases

Multi-IF — Results

Rank	Model	Score
1	lfm2-5-2-6b	80.07
2	lfm2-5-vl-3b	59.4
3	north-micro-vision-instruct	37.3

M

Multi-IF — Phases

1 phaseActive

Phase 1·

Multi-IF

Aug 4, 2026

Current

About

Meta's multi-turn, multilingual instruction-following benchmark — 4,501 three-turn conversations across 8 languages, extending IFEval.

Pages

Phases: 1Active
Participants: 1
Submissions: 3
Last Modified: Aug 4, 2026