OpenBMB Releases Fully Open-Source 1B AI Model MiniCPM5-1B, Citing 17.9 Artificial Analysis Score
OpenBMB and ModelScope released MiniCPM5-1B as a fully open-source 1 billion-parameter AI model, publishing its weights, training data and deployment code. The developers said it ranked first on Artificial Analysis among open models under 2 billion parameters with 17.9 points, ahead of Qwen3.5-2B at 16.3.
OpenBMB said an INT4 version is about 0.5 gigabytes, allowing MiniCPM5-1B to run on phones, browsers, laptops and other edge devices, and said it can power fully offline applications without cloud infrastructure. The project also said the model outperformed Qwen3.5-0.8B and LFM2.5-1.2B-Thinking on knowledge, math, coding and tool use.
From the sources (5 posts)
@vanstriendanielRT @ModelScope2022: MiniCPM5-1B is now fully open source, including weights, training data, and deployment code. 🚀1B params, #1 on Artifici…
@huggingfaceRT @ModelScope2022: MiniCPM5-1B is now fully open source, including weights, training data, and deployment code. 🚀1B params, #1 on Artifici…
@adinayakupMiniCPM5-1B is an impressive release in the 1B class! @OpenBMB ✨ 1B - Apache 2.0 ✨ Hybrid reasoning with Think / No-Think modes ✨ 128K context ✨ Runs on CPU/Apple Silicon/GPU ✨ Strong eval result in the same size
@modelscope2022MiniCPM5-1B is now fully open source, including weights, training data, and deployment code. 🚀1B params, #1 on Artificial Analysis among all open models under 2B (17.9 pts). 🤖 Beats Qwen3.5-2B (16.3) at half the par
@openbmbMiniCPM5-1B is now live — the strongest open‑source base model under 2B.🚀🚀 🔥 Ranks #1 on the Artificial Analysis (AA) index for small models, scoring 17.9 to beat the 2B-scale Qwen3.5-2B (16.3). ⚡ Comprehensively surpasses Qwen3.5-0.8B and