MiniCPM5-2B Tops HuggingFace Trending With 53.9 Avg Score Across 34 Benchmarks
面壁智能 · wechat · 2026-09-10
MiniCPM (面壁智能) announced that MiniCPM5-2B has topped the HuggingFace Trending chart. In the AA Intelligence Index V4.1.1 evaluation, the model ranked first among open-source base models under 4B parameters, averaging 53.9 across 34 benchmarks.
It targets on-device agent capabilities — tool calling, deep search, and code generation — aiming to bring agents to phones, PCs, and in-car systems, with Day0 support on Intel, Rockchip, and Arm platforms plus mainstream inference/fine-tuning frameworks.
Beyond weights, the release open-sources code pretraining, Agent SFT, and RL datasets, along with the UltraX data refinement framework, the Meshy RL framework, and the JustRLII RL strategy for reproducibility and app development.
Related event: OpenBMB's MiniCPM5-2B Tops Sub-4B Open Models in Intelligence Index(4 posts)→
More from Models
- Together AI sits down with Meta to explain Muse Spark Open — togethercompute · 2026-10-27
- GPT-6.1 Terra hits ~60% on Agents' Last Exam at $1/task, 9x cheaper than Opus 5.5 — ChrisGPT · 2026-10-27
- ChatGPT keeps telling tech-support users to email 404 Media's Jason Koebler — petrusenko_max · 2026-10-27
- GPT-6.1 Terra called the awkward middle tier, priced at half of GPT-6.1 Sol — cedric_chee · 2026-10-27
- Mystery model "Terra" scores 48 on Artificial Analysis index with no public API — haider1 · 2026-10-27
- GPT-6.1 Terra (Max) Appears on Artificial Analysis Scoring 48 — theanonymousone · 2026-10-27