@misc{indiciae19d00097391e,
  title = {Can Frontier LLMs Match Natively Multimodal Embeddings? A Comparison on Hard-Negative Text-to-Image Retrieval},
  author = {Archan Dutta and Vyanktesh Kanungo},
  year = {2026},
  url = {https://arxiv.org/abs/2608.11343},
  note = {Source identifier: 2608.11343}
}