How’s it going?

SLicenses for existing assets

All assets are used in compliance with their respective licenses.

Model License Citation URL
Qwen3-4B-Instruct-2507 Apache 2.0 [30][30]An Yang, Anfeng Li, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chang Gao, Chengen Huang, Chenxu Lv, Chujie Zheng, Dayiheng Liu, Fan Zhou, Fei Huang, Feng Hu, Hao Ge, Haoran Wei, Huan Lin, Jialong Tang, Jian Yang, Jianhong Tu, Jianwei Zhang, Jianxin Yang, Jiaxi Yang, Jing Zhou, Jingren Zhou, Junyang Lin, Kai Dang, Keqin Bao, Kexin Yang, Le Yu, Lianghao Deng, Mei Li, Mingfeng Xue, Mingze Li, Pei Zhang, Peng Wang, Qin Zhu, Rui Men, Ruize Gao, Shixuan Liu, Shuang Luo, Tianhao Li, Tianyi Tang, Wenbiao Yin, Xingzhang Ren, Xinyu Wang, Xinyu Zhang, Xuancheng Ren, Yang Fan, Yang Su, Yichang Zhang, Yinger Zhang, Yu Wan, Yuqiong Liu, Zekun Wang, Zeyu Cui, Zhenru Zhang, Zhipeng Zhou, and Zihan Qiu. Qwen3 technical report. Technical report, 2025. URL https://arxiv.org/abs/2505.09388. huggingface.co/Qwen/Qwen3-4B-Instruct-2507
Qwen3-4B-Base Apache 2.0 [30][30]An Yang, Anfeng Li, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chang Gao, Chengen Huang, Chenxu Lv, Chujie Zheng, Dayiheng Liu, Fan Zhou, Fei Huang, Feng Hu, Hao Ge, Haoran Wei, Huan Lin, Jialong Tang, Jian Yang, Jianhong Tu, Jianwei Zhang, Jianxin Yang, Jiaxi Yang, Jing Zhou, Jingren Zhou, Junyang Lin, Kai Dang, Keqin Bao, Kexin Yang, Le Yu, Lianghao Deng, Mei Li, Mingfeng Xue, Mingze Li, Pei Zhang, Peng Wang, Qin Zhu, Rui Men, Ruize Gao, Shixuan Liu, Shuang Luo, Tianhao Li, Tianyi Tang, Wenbiao Yin, Xingzhang Ren, Xinyu Wang, Xinyu Zhang, Xuancheng Ren, Yang Fan, Yang Su, Yichang Zhang, Yinger Zhang, Yu Wan, Yuqiong Liu, Zekun Wang, Zeyu Cui, Zhenru Zhang, Zhipeng Zhou, and Zihan Qiu. Qwen3 technical report. Technical report, 2025. URL https://arxiv.org/abs/2505.09388. huggingface.co/Qwen/Qwen3-4B-Base
Qwen3-8B Apache 2.0 [30][30]An Yang, Anfeng Li, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chang Gao, Chengen Huang, Chenxu Lv, Chujie Zheng, Dayiheng Liu, Fan Zhou, Fei Huang, Feng Hu, Hao Ge, Haoran Wei, Huan Lin, Jialong Tang, Jian Yang, Jianhong Tu, Jianwei Zhang, Jianxin Yang, Jiaxi Yang, Jing Zhou, Jingren Zhou, Junyang Lin, Kai Dang, Keqin Bao, Kexin Yang, Le Yu, Lianghao Deng, Mei Li, Mingfeng Xue, Mingze Li, Pei Zhang, Peng Wang, Qin Zhu, Rui Men, Ruize Gao, Shixuan Liu, Shuang Luo, Tianhao Li, Tianyi Tang, Wenbiao Yin, Xingzhang Ren, Xinyu Wang, Xinyu Zhang, Xuancheng Ren, Yang Fan, Yang Su, Yichang Zhang, Yinger Zhang, Yu Wan, Yuqiong Liu, Zekun Wang, Zeyu Cui, Zhenru Zhang, Zhipeng Zhou, and Zihan Qiu. Qwen3 technical report. Technical report, 2025. URL https://arxiv.org/abs/2505.09388. huggingface.co/Qwen/Qwen3-8B
GPT-OSS-20B Apache 2.0†\dagger [22][22]OpenAI. gpt-oss-120b & gpt-oss-20b model card. Technical report, OpenAI, 2025. arXiv:2508.10925. huggingface.co/openai/gpt-oss-20b
Gemini 3.1 Flash Lite Preview Proprietary API — ai.google.dev/gemini-api/terms
Table 30. Language models used as starting checkpoints for maze training or as LLM judges.
Dataset License Citation URL
SimpleQA-Verified MIT [10][10]Lukas Haas, Gal Yona, Giovanni D’Antonio, Sasha Goldshtein, and Dipanjan Das. Simpleqa verified: A reliable factuality benchmark to measure parametric knowledge, 2026. URL https://arxiv.org/abs/2509.07968. huggingface.co/datasets/google/simpleqa-verified
GSM8K MIT [7][7]Karl Cobbe, Vineet Kosaraju, Mohammad Bavarian, Mark Chen, Heewoo Jun, Lukasz Kaiser, Matthias Plappert, Jerry Tworek, Jacob Hilton, Reiichiro Nakano, Christopher Hesse, and John Schulman. Training verifiers to solve math word problems, 2021. URL https://arxiv.org/abs/2110.14168. github.com/openai/grade-school-math
MMLU MIT [11][11]Dan Hendrycks, Collin Burns, Steven Basart, Andy Zou, Mantas Mazeika, Dawn Song, and Jacob Steinhardt. Measuring massive multitask language understanding. In International Conference on Learning Representations, 2021. URL https://openreview.net/forum?id=d7KBjmI3GmQ. github.com/hendrycks/test
OR-Bench CC BY 4.0 [8][8]Justin Cui, Wei-Lin Chiang, Ion Stoica, and Cho-Jui Hsieh. OR-bench: An over-refusal benchmark for large language models. In Forty-second International Conference on Machine Learning, 2025. URL https://openreview.net/forum?id=CdFnEu0JZV. huggingface.co/datasets/bench-llm/or-bench
IMDB-CAD‡\ddagger Apache 2.0 [15][15]Divyansh Kaushik, Eduard Hovy, and Zachary Lipton. Learning the difference that makes a difference with counterfactually-augmented data. In International Conference on Learning Representations, 2020. URL https://openreview.net/forum?id=Sklgs0NFvr. github.com/acmi-lab/counterfactually-augmented-data
Table 31. Datasets and benchmarks used for evaluation or for sentiment-vector extraction.
Asset Version License Type URL
PyTorch 2.5.1, 2.7 BSD-style library github.com/pytorch/pytorch
HuggingFace Transformers 4.50 Apache 2.0 library github.com/huggingface/transformers
vLLM 0.6 Apache 2.0 library github.com/vllm-project/vllm
PEFT 0.13 Apache 2.0 library github.com/huggingface/peft
Flash-Attention 2 / 3 BSD-3-Clause library github.com/Dao-AILab/flash-attention
Modal — Proprietary service modal.com/legal/terms
Tinker API — Proprietary service thinkingmachines.ai/legal/tos/
Google Gemini API — Proprietary service ai.google.dev/gemini-api/terms
Table 32. Open-source code libraries and proprietary services used in training, inference, and evaluation.