Search papers, labs, and topics across Lattice.
This study conducts a comprehensive cost-benefit analysis comparing ten large language models (LLMs) and 26 embedding models across various tasks, revealing that while LLMs excel in reasoning-heavy retrieval, embedding models outperform in classification tasks. The findings indicate that the best LLM, Gemini 3.1 Pro, achieves a marginally higher performance than the top embedding model, but at a significantly higher cost鈥攗p to 1,431 times more expensive per benchmark pass. The research advocates for a strategic division of labor, recommending the use of embedding models for tasks like similarity and classification, while reserving LLMs for more complex reasoning tasks.
LLMs may outperform embedding models in reasoning, but the cost difference is staggering鈥攗p to 1,431 times more expensive for marginal gains.
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.
