Tech Meridian ← ENTITY INDEX
PROMY MERIDIAN RU

TOPIC · ENTITY #9264

mixture-of-experts

Related event timeline, sources and context from the news index.

EVENT TIMELINE

3

MODELS · 1 SOURCE · The Decoder

Aleph Alpha releases Kolibri, a 78B German–English MoE model with 1M-token context and open weights

Aleph Alpha has released Kolibri, a 78-billion-parameter German–English mixture-of-experts model that activates about 3 billion parameters per token, supports context windows up to one million tokens, and was trained on 768 B200 GPUs in Germany and Finland. The company says German makes up 21.3% of the training data, that Chinese models were used to generate synthetic data, and that Kolibri sits on the Pareto front of quality and operating cost; the weights are published under an Apache 2.0 license on Hugging Face and the model is positioned for public administration, aviation, and industry with EU AI Act compliance in mind.

8.0

RESEARCH · 1 SOURCE · arXiv cs.AI

CourseChat: on‑premises multi-course RAG tutor for business education (arXiv:2610.02510v1)

The authors present CourseChat, an on-premises, multi-course retrieval-augmented generation (RAG) tutor deployed behind a campus web gateway and integrated for use with Moodle; it runs twin edge hosts with FastAPI, a local vector database, and an LLM served via Ollama across six isolated course CRNs. The paper reports two model bake-offs, a source-fidelity comparison, conversation and quiz audits, finds that larger models failed classroom speed constraints while 12B and 7B passed and a mixture-of-experts introduced new errors, and retains an 8B production model pending demonstrated overall improvement; the study does not evaluate learning gains or public-gateway acceptance.

6.0

RESEARCH · 1 SOURCE · The Decoder

Study: humans make most decisions when AI agents help develop Atria Dawn Preview

Researchers from Fudan University analyzed 769 task and agent logs from development of the agentic LLM Atria Dawn Preview (a 744B mixture-of-experts model). They found AI provided many proposals and increased the number of agent actions, but humans made about 85% of methodological decisions and 93% of content-related decisions; roughly one-third of tasks were judged infeasible without AI.

7.0