Haotian Liu

Researched 16d ago

10 models across 3 families · Latest: LLaVA 1.6 Vicuna 7B (2024-01)

Academic researcher focused on vision models

VisionClassificationJSON / Tool useIndividual

Haotian Liu's portfolio covers 8 active models across 3 non-obsolete families, with task labels spanning vision, classification, and json / tool use. Open a model detail page to compare provider routes and sourced benchmarks.
Portfolio context: 3 decision-task tags, 8 active tracked models, latest research stamp 2026-05-19.

Use this portfolio page for

Teams evaluating vision, classification, and json / tool use across this lab's releases
Readers comparing families before locking a flagship SKU
8 tracked SKUs for migration and pricing follow-ups

Do not stop here for

Choosing a hosting provider without opening a model page for price ladders

Active models

Non-deprecated SKUs linked to this researcher

Active families

Non-obsolete families in coverage

Open catalog

1 OSS

0 open-weight (text match)

Decision task tags

Mapped to the site-wide task taxonomy

Latest dated release

2024-01-31

LLaVA 1.6 Vicuna 7B

Freshness

2026-05-19

Researched 16d ago

fresh

Release cadence

Showing 5 recent dated ships (full timeline below). Latest spotlight: LLaVA 1.6 Vicuna 7B (2024-01-31).

Where this lab wins

Vision: 3 tracked models with multimodal benchmark coverage.
Classification: 1 tracked model with MMLU-class moderation/safety coverage.
JSON/tool-use: 1 tracked model with BFCL / Nexus strict-JSON routing coverage.

Flagship quality / price signal

Anchor SKU: LLaVA Vicuna 13B (best sourced coding Q/$ in this portfolio).

Quality / dollar unavailable for this anchor — missing benchmark coverage and/or output token price on the cheapest ladder route (open the model detail after pricing lands).

Haotian Liu is an AI research organization founded in N/A. Academic researcher focused on vision models. Haotian Liu ships 3 model families totaling 10 models, with the most recent release LLaVA 1.6 Vicuna 7B in 2024-01. Notable families include LLaVA 1.6, LLaVA 1.5, and LLaVA. Use it as a stable reference for lab background, release coverage, and follow-up model pages as they. View official API endpoints, benchmark performance, and coding/agent fit for every Haotian Liu model.

About

Haotian Liu is a distinguished AI researcher who has made substantial contributions to the fields of generative AI and large language models (LLMs). His academic path was marked by rigorous training and research in computer vision and machine learning, commencing with a bachelor's degree at Zhejiang University. Liu advanced this knowledge during his Ph.D. at the University of Wisconsin-Madison, under the mentorship of Professor Yong Jae Lee, where he honed his expertise in integrating visual and textual data. A hallmark of Liu's research is the development of the Large Language and Vision Assistant (LLaVA). This cutting-edge platform employs sophisticated visual instruction tuning methods to boost the performance of LLMs, allowing them to interpret and generate responses based on visual input effectively. His work aims for these AI systems to rival the comprehension level of models like GPT-4, thereby facilitating complex interactions that merge linguistic and visual elements. This advancement is critical as it enhances the AI’s contextual understanding and reasoning capabilities, which are vital for applications in diverse sectors, including biomedical research and education. In his pursuit of applying AI to specialized fields, Liu spearheaded the creation of LLaVA-Med, a variant of LLaVA designed specifically for biomedical use. This model harnesses large-scale datasets derived from PubMed Central, enhancing its ability to address intricate biomedical image-related inquiries. A notable feature of Liu’s methodology is the use of a curriculum learning strategy that allows the model to adapt from simpler tasks to more complex ones, emulating human learning processes in biomedical sciences. This innovative approach underscores Liu’s dedication to developing accessible, domain-centric AI tools for professionals. Beyond these critical projects, Liu's research portfolio includes papers focused on applying generative AI to practical challenges, particularly the fusion of multiple modalities for real-world applications such as visual question answering and image captioning. His endeavors emphasize bridging the gap between language and image understanding, aiming to forge AI systems that respond precisely to intricate human queries with contextual relevance. Haotian Liu’s work is characterized by a profound emphasis on multimodal learning and the personalization of AI systems to align with user needs. His pioneering methods are transforming AI into a more intelligent and interactive tool that can support professionals across various domains, especially where detailed domain knowledge is required. As the field of generative AI continues to expand, Liu’s efforts are instrumental in steering the future of AI development both in academia and applied settings.