Alternatives to Transformers

Open-source library for parameter-efficient fine-tuning of transformer models.

Open-weight multilingual AI with hybrid Transformer-Mamba architecture from TII.

Korean long-context BigBird model for fill-mask and fine-tuning up to 4096 tokens

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

Frontier code models with 5M-token context, built to automate software engineering and research.

Open-source Python library for guaranteed valid structured outputs from LLMs

Fast Java tokenizer library for OpenAI GPT models using BPE

StableLM is an open-source, self-hostable LLM suite from Stability AI for transparent text and code generation, with 3B and 7B Alpha models under permissive

Open-source bilingual Chinese-English 7B LLM for research and fine-tuning

Open-source omni-modal LLM for personalized, long-horizon speech with vision and text input.

Google's API for text generation with PaLM and Gemini models, plus MakerSuite for prototyping.

Reka ships omni models and edge AI for real-time video reasoning on devices.

Fine-tune and deploy open-source LLMs with managed GPU infrastructure.

Chinese enterprise AI platform with open-source GLM models, MaaS APIs, and autonomous agents

Open-source MoE vision-language model for efficient agent inference

Zhipu AI's GLM-5.2 open-source coding model with 1M context and autonomous agents for Chinese enterprises.

Mistral: European frontier AI platform for GDPR-compliant agents, custom models, and sovereign deployment.

Pythonic constrained LLM generation with regex and CFGs

Free, community-maintained GitHub repo of OpenAI API code examples and guides for Python developers.

Production-grade multilingual NLP toolkit with 300+ Chinese/English models and 104 language support.

Open-weight on-device AI models for private, low-latency edge intelligence—free to use under $10M revenue.

Open-source text-to-video model from a two-person lab, Apache 2.0, self-hostable.

Open-source instruction-following LLM you can fine-tune on a single GPU

Open-source ultra-fast small language models for edge inference

Open-source attention-sink library for constant-memory, endless LLM chat generation

State-of-the-art tokenization library: train custom vocabularies with BPE, WordPiece, Unigram—free and open source.

A single API for semantic search and LLM text generation.

Open-source frontier models and agent-first office doc tooling for specialized knowledge work.

Sovereign specialized LLMs for European enterprise AI on EU infrastructure.

Chinese enterprise LLM family with a medical AI focus