Nanbeige and LFM2.5 tie for top mobile model score of 63
ArtificialAnlys · x · 2026-08-25
At the standard 16K context limit, Nanbeige4.2-3B (Reasoning) and LFM2.5-2.6B (Reasoning) tie for the top average score at 63. Individual highlights include Qwen3.5 9B leading in BFCL and GPQA Diamond, Falcon-H1R-7B achieving 97% on MATH-500, and LFM2.5-2.6B showing significantly lower hallucination rates. Nanbeige4.2-3B won no single evaluation but ranked top-five in BFCL, GPQA Diamond, and MATH-500, securing its overall tie.
Related event: Artificial Analysis and Liquid AI Launch On-Device Small Model Benchmark(9 posts)→
More from Models
- Zuck announces Meta Muse Spark Open weights: 816B-param image-text model tuned for Mac Studio — bansalg_ · 2026-10-16
- Google launches revamped Gemma docs hub covering fine-tuning, inference and model internals — _philschmid · 2026-10-16
- Post-trained NVIDIA Nemotron-3.5-Super-VL hits 35% on AutoResearchExam, matching a 20× larger model — AlexGDimakis · 2026-10-16
- Nathan Lambert slams Meta's Muse Open License as 'trash' — natolambert · 2026-10-16
- Google launches Gemini 4 Argon to Ultra and Pro subscribers, touting physics and 3D skills — ShikharMurty · 2026-10-16
- Google gives X users early access to test Argon, not just companies — lyraxana · 2026-10-16