Alternatives to Yi

Enterprise AI decision platform for top-down transformation in China, led by CEO mandate.

Zhipu GLM delivers open-source LLM models, MaaS APIs, and autonomous agents for Chinese enterprises and developers.

Chinese enterprise LLM family with a medical AI focus

Zhipu AI's GLM-5.3 and MaaS platform deliver open-source SOTA coding with 1M context for enterprise agents.

Enterprise AI cost optimizer: route models and cut token waste

Korean enterprise LLM — consumer CLOVA X service ended April 2026

AI research lab building edge-first omni models that reason over video, vision, and text in real time.

Sovereign specialized LLMs: Aleph Alpha Pharia trains custom domain models on EU infrastructure for regulated European organizations.

Open-source 198B-A11B MoE vision-language model for efficient agent inference

AI-native cloud for high-throughput open-source model inference and GPU compute at scale.

Mistral delivers sovereign frontier AI models, agents, and OCR with GDPR-compliant EU and self-hosted deployment.

Apache 2.0 open-weight LLM family with hybrid Transformer-Mamba architecture

AI-native CDP that unifies customer data and automates AI-driven segmentation and activation.

Open-source multimodal model with native tool use for building autonomous agents that see and act.

Open-weight on-device AI with native audio, vision, and Japanese models, free under $10M revenue.

Free bilingual Chinese-English 7B base LLM for research and fine-tuning on Hugging Face

Databricks' open-source instruction-following LLM for fine-tuning on a single GPU, free and commercially licensed.

Open-source Dutch BERT model with state-of-the-art performance for Dutch NLP.

The open-source Python library for loading, fine-tuning, and running transformer models across text, vision, audio, and video.

An open-source small language model kit for fast local inference on modest hardware.

StableLM: Stability AI's open-source language models for transparent, self-hosted text and code generation

Google's PaLM API lets developers build with Google's language models, now including Gemini.

Free Korean BERT pretrained on 12.5GB of noisy Naver news comments for informal text understanding.