My Name is Riemann: staged open model for English-Arabic code and reasoning
A 13.1B-parameter prerelease language model for English and Arabic text generation, code completion and explanation, and step-by-step problem solving, built by expanding the smaller OpenPeerLLM model in stages and carrying weights forward, with a small controller that adjusts layer outputs without changing base weights.
Why it matters
Its staged-expansion training is the experiment: instead of training each size from scratch, weights carry forward across six stages, and the repo ships the weights, controller, training pipeline and docs for others to inspect.
What you could build with it
A researcher studying efficient training could reproduce one expansion stage from the published pipeline: start from the smaller checkpoint, expand, and measure whether carried-forward weights beat a from-scratch run of the same size on code and bilingual tasks.
Does it hold up?
Too early to judge: an experimental prerelease without instruction or safety tuning and no capability benchmarks; the repo's full pipeline and docs make it inspectable, but production use would need further training and evaluation first.
Built with My Name is Riemann
- Riemann Computing's AI model prerelease surpasses 500 downloads on Hugging Facearticle · October 28 announcement of the prerelease milestone, the staged-expansion method, the controller-based fine-tuning approach, and the caveat that it lacks instruction and safety tuning.
Learn more
First spotted on hf: source.
More AI models
Tencent Hunyuan Hy3: open 295B MoE with 21B active under Apache 2.0Tencent open-sourced Hy3, a 295B-parameter mixture-of-experts model with only 21B parameters active per token, under…models · JEV 0.73NVIDIA Isaac Lab 3.0 GA + Isaac OS open runtime + Agile ONE SN humanoidMajor Isaac platform expansion at GTC Berlin Oct 21: Isaac Lab 3.0 is generally available as the open-source robot…models · JEV 0.73Kimi K3.1 Pro: Moonshot's 2.8T open-weight flagship officially launchesMoonshot AI officially launched Kimi K3.1 Pro on Oct 23, a mixture-of-experts model with 2.8 trillion total / 104…models · JEV 0.72NVIDIA Nemotron 3.5 Retriever collection: Embed 8B/1B, Rerank 8B, Parse 631MNVIDIA's Oct 21 collection drop for agentic retrieval: Nemotron-3.5-Embed in 8B BF16, 1B BF16 and 1B NVFP4…models · JEV 0.72NVIDIA Cosmos 3.5 world models: now openly availableNVIDIA's Cosmos 3.5 family of open world foundation models for robots, autonomous vehicles and vision agents is now…models · JEV 0.71NVIDIA RFD4-Proteina: open protein-design BioNeMo modelOpen protein-design model from NVIDIA and the Baker Lab at UW Medicine's Institute for Protein Design, announced at…models · JEV 0.7
Get the week's best AI launches, plus 3 ideas worth building
One email every Saturday. Ranked by traction, not hype. Free.