ELYZA Open Models: Model List and Licenses
About ELYZA
According to its organization card, ELYZA is an AI company spun out of the Matsuo Laboratory at the University of Tokyo, with the mission of “creating the new normal in unexplored territory". It conducts joint research and product development toward putting large language models into practical use, and releases Japanese models on its elyza organization on Hugging Face.
What It Releases
- ELYZA-japanese-Llama-2 (2023): 7B and 13B models built by further training Llama 2 for Japanese, including fast versions with an expanded Japanese vocabulary.
- Llama-3-ELYZA-JP-8B (June 2024): Llama 3 8B adapted for Japanese, with official GGUF and AWQ builds.
- ELYZA-Thinking-1.0-Qwen-32B / ELYZA-Shortcut-1.0 (April 2025): models based on Qwen2.5-32B-Instruct with stronger Japanese reasoning. According to the model card, they were trained by imitation learning on synthetic data with long chains of thought generated by a Monte Carlo Tree Search (MCTS) based algorithm.
- ELYZA-Diffusion-1.0 (December 2025): a Japanese diffusion language model, based on the diffusion LLM Dream-v0-7B and further trained on large-scale Japanese data. It generates text by iteratively denoising from an all-MASK sequence.
Licenses
They depend on the base model. The models based on Qwen2.5 and Dream (Thinking, Shortcut, Diffusion) are Apache-2.0, and the Thinking model card states “Built with Qwen". The models based on Llama 3 and Llama 2 use the respective Llama licenses. Read each model’s license text before use. This summary is not legal advice.
Running Them Locally
- The only official GGUF is Llama-3-ELYZA-JP-8B (as of 2026-09-26); for the others, look for third-party conversions.
- The diffusion language model (ELYZA-Diffusion) generates text differently from ordinary language models. llama.cpp and Ollama may not support it.
Sources: the elyza organization card; model cards of elyza/ELYZA-Thinking-1.0-Qwen-32B and elyza/ELYZA-Diffusion-Instruct-1.0-Dream-7B; the model list of the elyza organization on Hugging Face (all as of 2026-09-26).
Our Coverage and Data
Local Model Watch has not published an article on ELYZA’s models yet. The model list below is collected by code from Hugging Face every day. Part of our publisher index; model families link to their family pages.
At a Glance
| Item | Value |
|---|---|
| Hugging Face organization | elyza |
| Models on Hugging Face | 18 |
| Articles on its own models | 0 |
| Articles on derived models | 0 |
Recent Models on Hugging Face
The publisher lists 18 models in total on Hugging Face. Below are up to 15 of the most recent generative models (text, image, video, audio and so on), the scope of this site, per organization. Classifiers, feature extractors, research components and repositories without a model card are omitted, and the publisher’s own quantized builds and helper variants are folded into the row of the original model. Parameter counts are left blank for repositories that pack weights into integer types, where Hugging Face’s count is not the real parameter count.
→ Scroll horizontally to see all columns
| Released | Model | Task | Parameters | License | Official variants |
|---|---|---|---|---|---|
| 2025-12-10 | elyza/ELYZA-Diffusion-Base-1.0-Dream-7B | — | 7.6B | apache-2.0 |
— |
| 2025-12-10 | elyza/ELYZA-Diffusion-Instruct-1.0-Dream-7B | — | 7.6B | apache-2.0 |
— |
| 2025-05-01 | elyza/ELYZA-Shortcut-1.0-Qwen-7B | text generation | 7.6B | apache-2.0 |
— |
| 2025-05-01 | elyza/ELYZA-Shortcut-1.0-Qwen-32B | text generation | 32.8B | apache-2.0 |
— |
| 2025-04-30 | elyza/ELYZA-Thinking-1.0-Qwen-32B | text generation | 32.8B | apache-2.0 |
— |
| 2024-06-25 | elyza/Llama-3-ELYZA-JP-8B | text generation | 8.0B | llama3 |
GGUF, AWQ |
| 2023-12-26 | elyza/ELYZA-japanese-Llama-2-13b-fast | text generation | — | llama2 |
instruct |
| 2023-12-26 | elyza/ELYZA-japanese-Llama-2-13b | text generation | — | llama2 |
instruct |
| 2023-11-07 | elyza/ELYZA-japanese-CodeLlama-7b | text generation | 6.7B | llama2 |
instruct |
| 2023-08-28 | elyza/ELYZA-japanese-Llama-2-7b-fast | text generation | — | llama2 |
instruct |
| 2023-08-28 | elyza/ELYZA-japanese-Llama-2-7b | text generation | — | llama2 |
instruct |
From the Hugging Face API. Last changed 2026-09-27 (JST). Collected by code.
Last updated 2026-09-27 (JST). The explanation at the top of this page was written with the help of AI from the primary sources it cites. The tables and lists under “Our Coverage and Data" are assembled by code from our article log and the Hugging Face API.