Gespeichert in:
| Hauptverfasser: | , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | https://arxiv.org/abs/2605.10391 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866914553571311616 |
|---|---|
| author | Phoenix, Team : Ray, Arka Jawad, Askar Ali Mohamed Lee, Biondi Seah, Elijah Lim, Eva Teo, Fiona Toh, Grace Teo, Guang Xiang Tan, Jun En Bong, Jia Hui Wang, Jiale Ng, Jonathan Tan, Justin Yew, Kai Zhe Ong, Matthew Yeo, Shun Yi Lam, Wen Jett Tan, Wen Xiu Zhang, Ze Yu Ng, Gee Wah Ang, Chee Wee AI, Mistral : Sadé, Adrien Kunsch, Guillaume Loh, Jia Sin Schuhl, Nicolas Menneer, Rupert Jamil, Umar Maladière, Vincent Pan, Yimu |
| author_facet | Phoenix, Team : Ray, Arka Jawad, Askar Ali Mohamed Lee, Biondi Seah, Elijah Lim, Eva Teo, Fiona Toh, Grace Teo, Guang Xiang Tan, Jun En Bong, Jia Hui Wang, Jiale Ng, Jonathan Tan, Justin Yew, Kai Zhe Ong, Matthew Yeo, Shun Yi Lam, Wen Jett Tan, Wen Xiu Zhang, Ze Yu Ng, Gee Wah Ang, Chee Wee AI, Mistral : Sadé, Adrien Kunsch, Guillaume Loh, Jia Sin Schuhl, Nicolas Menneer, Rupert Jamil, Umar Maladière, Vincent Pan, Yimu |
| contents | We introduce Phoenix-VL 1.5 Medium, a 123B-parameter natively multimodal and multilingual foundation model, adapted to regional languages and the Singapore context. Developed as a sovereign AI asset, it demonstrates that deep domain adaptation can be achieved with minimal degradation to broad-spectrum intelligence and alignment. Continued pretraining was performed on Mistral Medium 3.1 using a localized 1-trillion tokens multimodal corpus, followed by a 250-billion tokens long-context extension phase. Subsequent post-training incorporated a novel human-annotated Singapore multimodal dataset and curated textual corpus on Singapore culture, knowledge, and legislation, totaling 22-billion tokens. An additional 5 billion tokens of model alignment was performed through Online Direct Preference Optimization. Phoenix-VL 1.5 Medium achieves state-of-the-art performance for its size on Singapore multimodal, legal, and government policy benchmarks while remaining globally competitive on general multimodal intelligence, multilingual, and STEM benchmarks. We also introduce a novel evaluation suite encompassing localized knowledge benchmarks and an institutionally aligned model behavior and safety framework. We report the data curation principles, training methodology, and highlight benchmark and inference performance. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_10391 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Phoenix-VL 1.5 Medium Technical Report Phoenix, Team : Ray, Arka Jawad, Askar Ali Mohamed Lee, Biondi Seah, Elijah Lim, Eva Teo, Fiona Toh, Grace Teo, Guang Xiang Tan, Jun En Bong, Jia Hui Wang, Jiale Ng, Jonathan Tan, Justin Yew, Kai Zhe Ong, Matthew Yeo, Shun Yi Lam, Wen Jett Tan, Wen Xiu Zhang, Ze Yu Ng, Gee Wah Ang, Chee Wee AI, Mistral : Sadé, Adrien Kunsch, Guillaume Loh, Jia Sin Schuhl, Nicolas Menneer, Rupert Jamil, Umar Maladière, Vincent Pan, Yimu Computation and Language Artificial Intelligence Computer Vision and Pattern Recognition We introduce Phoenix-VL 1.5 Medium, a 123B-parameter natively multimodal and multilingual foundation model, adapted to regional languages and the Singapore context. Developed as a sovereign AI asset, it demonstrates that deep domain adaptation can be achieved with minimal degradation to broad-spectrum intelligence and alignment. Continued pretraining was performed on Mistral Medium 3.1 using a localized 1-trillion tokens multimodal corpus, followed by a 250-billion tokens long-context extension phase. Subsequent post-training incorporated a novel human-annotated Singapore multimodal dataset and curated textual corpus on Singapore culture, knowledge, and legislation, totaling 22-billion tokens. An additional 5 billion tokens of model alignment was performed through Online Direct Preference Optimization. Phoenix-VL 1.5 Medium achieves state-of-the-art performance for its size on Singapore multimodal, legal, and government policy benchmarks while remaining globally competitive on general multimodal intelligence, multilingual, and STEM benchmarks. We also introduce a novel evaluation suite encompassing localized knowledge benchmarks and an institutionally aligned model behavior and safety framework. We report the data curation principles, training methodology, and highlight benchmark and inference performance. |
| title | Phoenix-VL 1.5 Medium Technical Report |
| topic | Computation and Language Artificial Intelligence Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2605.10391 |