IFBench (AllenAI)

Out-of-distribution precise instruction following.

The best 3 · 3 models tested

Top 3 · best setting per model · 3 models, 4 results (0 independent)

Measured by independent testersIndependent Reported by the makerVendor-reported

IFBench (AllenAI) leaderboard
				Thinking levelSetting
1	Muse Glimmer 30B	Meta (Meta Superintelligence Labs, Muse)	77.0%	HighHigh reasoning	Maker's own figureVendor-reportedMeta ↗	—	10 Aug 2026
2	Mistral Medium 3.5	Mistral AI	69.0%	Highmaximum reasoning settings (reasoning_effort=high)	Maker's own figureVendor-reportedMistral AI ↗	—	22 May 2026
3	Mistral Small 4	Mistral AI	48.0%	HighReasoning (reasoning_effort=high)	Maker's own figureVendor-reportedMistral AI ↗	—	16 Mar 2026