Mid-size: 10 to 70 billion
The workhorse open models: Llama 2 13B, Qwen 32B, Gemma 27B and their kin. 103 models.
馃嚭馃嚫 United StatesTalkie
A 13B small language model trained only on pre-1931 public domain English textSmall 路 Open weights 路 2026
馃嚭馃嚫 United StatesPythia
A suite of 16 models for interpretability research.Research 路 Open weights 路 2023
WorldINTELLECT-1
A 10 billion parameter model trained by a global community.General 路 Open weights 路 2024
馃嚚馃嚦 ChinaDeepSeekMoE
Mixture-of-experts language model architecture by DeepSeek AIGeneral 路 Open weights 路 2024
馃嚝馃嚪 FranceT0
Encoder-decoder model for zero-shot task generalisation via promptsGeneral 路 Open weights 路 2021
馃嚭馃嚫 United StatesGPT-NeoX-20B
Open 20 billion parameter autoregressive language model by EleutherAIGeneral 路 Open weights 路 2022
馃嚚馃嚦 ChinaCogVLM
Open-source visual language model with deep feature fusionMultimodal 路 Open weights 路 2023
馃嚭馃嚫 United StatesCerebras-GPT
Open compute-optimal language models from Cerebras SystemsGeneral 路 Open weights 路 2023