ModelBest Releases On-Device Model MiniCPM5-2B
面壁智能 · wechat · 2026-07-19
During WAIC, ModelBest released the on-device model MiniCPM5-2B, featuring 2B parameters, hybrid thinking, a 512K ultra-long context, and local deployment capabilities for smartphones, PCs, and smart cockpits.
Core selling points include achieving leading results among sub-4B models on benchmarks like AA-Index/ArtificialAnalysis, with standout performances in long-text understanding, Agent capabilities, and mathematical reasoning. The release also introduced the UltraX data governance solution and an end-to-end training pipeline for on-device agents.
On the ecosystem front, ModelBest partnered with the FlagOS community to achieve multi-chip Day 0 adaptation, covering platforms like Ascend, Hygon, Kunlunxin, Metax, Moore Threads, T-Head, Tsingmicro, Iluvatar CoreX, and Nvidia. They plan to release adaptation images and deployment guides alongside the open-source model.
More from Infra
- SmolVM open-sources persistent computer infrastructure for agents that outlive chat sessions — aniketmaurya · 2026-09-11
- PyTorch Day Korea 2026 launches first offline conf, CFP closes Sept 13 — PyTorch · 2026-09-11
- Local LLM server dilemma: 4x CMP-170HX (price up 53% in 20 days) vs Mac Studio M5 Ultra — rumboll · 2026-09-11
- llama.cpp lands Flash Attention tuning for RDNA4, big prefill gains on AMD — pmttyji · 2026-09-11
- Your p99 latency benchmark may be lying: a deep dive into coordinated omission — Franc0Fernand0 · 2026-09-11
- Running MiniMax H3 on 12GB VRAM: quantization, Turbo LoRAs and attention backends compared — Possible_Mood676 · 2026-09-11