Llama 4 Maverick holds 85.5% MMLU lead among open-weight models — Meta's continued open-source posture vindicated against 2025 analyst predictions of a closed pivot
Llama 4 Maverick's 85.5% MMLU lead among open models gives Meta a clear win after a year of speculation that Llama 5 would go closed-source. The result counters the narrative that only Chinese labs are pushing open weights at the frontier and re-centers Meta in the open-source frontier conversation through H2 2026.
The substantive piece is the open-source-posture vindication. Through Q4 2025 and Q1 2026, multiple analyst frameworks predicted Meta would close-source Llama 5 in response to commercial-licensing economics and frontier-lab competitive pressure. Maverick's continued open-weight release with a clear MMLU-leadership position among open models demonstrates that Meta's calculated bet on open-source-as-strategic-asset is producing the intended competitive result; the closed-pivot prediction was structurally wrong.
The structural read against Qwen 3.5's multilingual open-source default position is that the English-primary and multilingual open-source procurement segments now have clear category leaders that aren't from the same vendor. Two-vendor segmentation of the open-source frontier is more durable than single-vendor consolidation would have been; the H2 2026 open-source frontier-model competitive landscape has multiple winners across structurally-different procurement segments.
Featherless — Best Open-Source LLMs in 2026 → · Medium — The Free Models Running the World → · Till Freitag — Open-Source LLM Comparison 2026 →