01.AI

weights

open(Apache-2.0, on HF)

data

closed(pretraining corpus not released)

code

partial(inference/usage examples

license

Apache-2.0(OSI, permissive, no use restrictions)

Permissive Apache-2.0 weights, but the training data and the full training code are not released, so this is open weights rather than open source.

• https://api.github.com/repos/01-ai/Yi-1.5 recorded 2026-08-13

  `private: false`, `archived: false`, license `Apache-2.0`, 561 stars, last pushed 2024-11-11. Usage examples only, no training pipeline, which is what holds `code` at partial.

• https://cdn.jsdelivr.net/gh/01-ai/Yi-1.5@main/ recorded 2026-09-29

  jsDelivr listing of the 01-ai/Yi-1.5 GitHub repo at main. The top level holds only "/gh/01-ai/Yi-1.5@main/LICENSE", "/gh/01-ai/Yi-1.5@main/NOTICE", "/gh/01-ai/Yi-1.5@main/README.md", "/gh/01-ai/Yi-1.5@main/demo/" and "/gh/01-ai/Yi-1.5@main/requirements.txt", with no training or data-preparation code.

• https://cdn.jsdelivr.net/gh/01-ai/Yi-1.5@main/LICENSE recorded 2026-09-29

  LICENSE file in the Yi-1.5 repo (via jsDelivr) is the standard Apache License text, "TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION", with no added use restrictions.

• https://cdn.jsdelivr.net/gh/01-ai/Yi-1.5@main/README.md recorded 2026-09-29

  Yi-1.5 repo README (via jsDelivr). It covers inference, Ollama, vLLM and a web demo, and for training points to third-party tools - "These frameworks all support fine-tuning the Yi series models" - with no pretraining pipeline. It states "The code and weights of the Yi-1.5 series models are distributed under the" Apache 2.0 license.

• https://huggingface.co/01-ai/Yi-1.5-34B recorded 2026-09-29

  Yi-1.5-34B model page. The card says the model is "continuously pre-trained on Yi with a high-quality corpus of 500B tokens" without naming or linking any dataset, and the page carries the tag "license:apache-2.0". The page metadata records the repo as not gated; searched for agree to share and access request, none.

• https://huggingface.co/01-ai/Yi-1.5-9B recorded 2026-09-29

  Yi-1.5-9B model page, same card text, with the tag "license:apache-2.0" and the line "Yi-1.5-9B is the top performer among similarly sized open-source models" under base models. The page metadata records the repo as not gated.

• https://huggingface.co/api/models/01-ai/Yi-1.5-34B recorded 2026-09-29

  Hugging Face API record for 01-ai/Yi-1.5-34B lists the weight shard "model-00001-of-00015.safetensors" through 00015 and the tag "license:apache-2.0"; the gated and private fields are both false.

Adoption is measured as Hugging Face downloads for the single declared checkpoint, 01-ai/Yi-1.5-34B. Yi is a 2024-era family now well behind the Qwen, Llama and DeepSeek generations, and its downloads have fallen off accordingly.

• https://huggingface.co/api/models/01-ai/Yi-1.5-34B recorded 2026-10-07

  9,555 downloads in the trailing 30 days for 01-ai/Yi-1.5-34B

Yi-34B, the original 2023 base model, scores 76.3 on MMLU, 54.3 on BBH and 81.4 on C-Eval in its tech report, competitive for its time but well below current DeepSeek and Qwen releases and comparable to Falcon 3. The Yi-1.5 model cards show base-model benchmarks only as charts, so the published figures are for Yi-34B rather than Yi-1.5-34B.

• https://arxiv.org/abs/2403.04652 recorded 2026-09-29

  Yi foundation-models tech report abstract. It states "Our base models achieve strong performance on a wide range of benchmarks like MMLU" but carries none of the recorded figures.

• https://arxiv.org/html/2403.04652v1 recorded 2026-09-29

  Yi tech report full text. Table 2, "Overall performance on grouped academic benchmarks compared to open-source base models", gives the Yi 34B row as MMLU "76.3", BBH "54.3" and C-Eval "81.4". These are Yi-34B figures, not Yi-1.5-34B.

• https://cdn.jsdelivr.net/gh/01-ai/Yi-1.5@main/README.md recorded 2026-09-29

  Yi-1.5 repo README (via jsDelivr) carries no benchmark table; it says "For model details and benchmarks, see" the Hugging Face model card collection.

• https://github.com/01-ai/Yi-1.5 recorded 2026-08-13

  Cited for Yi-1.5 base benchmark results (MMLU 76.3, BBH 54.3, C-Eval 81.4), but the README no longer carries a benchmark table - it says "For model details and benchmarks, see Model Card", and none of the three figures appears on either base card. The recorded value has no live source here.

• https://huggingface.co/01-ai/Yi-1.5-34B/raw/main/README.md recorded 2026-09-29

  Yi-1.5-34B raw model card. Under base models it says "Yi-1.5-34B is on par with or excels beyond larger models in some benchmarks" and shows the results only as an image, with no benchmark figures in the text.