Search papers, labs, and topics across Lattice.
This study conducts a comprehensive cost-aware evaluation of ten large language models (LLMs) and 26 embedding models across 37 tasks, revealing that while LLMs excel in reasoning-heavy retrieval, embedding models outperform in classification tasks. The performance of the best LLM, Gemini 3.1 Pro, is only marginally better than the top embedding model, with a difference of just 0.4 points. However, the significant cost disparity鈥攗p to 1,431 times more expensive for LLMs鈥攕uggests a strategic division of labor, advocating for embedding models in similarity, classification, and clustering tasks, while reserving LLMs for more complex reasoning tasks.
LLMs may offer slight performance advantages, but their exorbitant costs and slower processing times make embedding models the smarter choice for most tasks.
Should you replace your text-embedding pipeline with a large language model? We answer this with a controlled, cost-aware comparison of ten LLMs across six families and 26 embedding models (118M to 14B parameters) on 37 tasks spanning classification, semantic textual similarity (STS), clustering, pair classification, and retrieval. In aggregate the two paradigms are effectively tied: the best LLM (Gemini 3.1 Pro, 77.6) and the best embedding model (77.2) differ by 0.4 points. Their strengths differ by task: LLMs lead on reasoning-heavy retrieval, embedding models lead on classification, and the two match on clustering, STS, and pair classification. Reaching that parity is expensive. An LLM costs up to 1,431x more than an embedding model of comparable quality (USD 154 vs. USD 0.11 per benchmark pass), and the open LLMs tested process tokens 2.5 to 736x more slowly on the same GPU. Reasoning tokens account for 28 to 81% of LLM inference cost; lower reasoning budgets preserve or improve retrieval quality for most models in our ablation. The Pareto frontier contains the leading embedding models and one LLM, Gemini 3.1 Pro. These results support a division of labour: use embedding models for similarity, classification, and clustering, and reserve LLMs for reasoning-intensive retrieval. Our code, datasets, and results are publicly available at https://github.com/embeddings-benchmark/embedders-dilemma.
