AI Paper Atlas

Research map / Architectures and efficiency / Quantization

TurboQuant: Online Vector Quantization with Near-optimal Distortion Rate

Amir Zandieh, Majid Daliri, Majid Hadian, Vahab Mirrokni

ICLR 2026 · Architectures and efficiency › Quantization · approach: Post-training weight quantization · 99 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
quantize KV cache with TurboQuant at 3.5 bitsLlama-3.1-8B-Instruct · LongBench-V1average score: 50.06 -> 50.06No clear effect
quantize KV cache with TurboQuant at 2.5 bitsLlama-3.1-8B-Instruct · LongBench-V1average score: 50.06 -> 49.44Hurt
TurboQuant 2.5 bit vs KIVI 3 bitKV cache quantization · LongBench-V1 Llama-3.1-8Baverage score: 48.50 -> 49.44Improved
TurboQuant 2.5 bits on Ministral-7B-InstructMinistral-7B-Instruct · LongBench-V1average score: 49.89 -> 49.62Hurt
use TurboQuant instead of Product Quantization / RabitQ for indexingvector quantization · NN search quantization time (d=1536)quantization time: 239.75 -> 0.0013Improved
add 1-bit QJL on MSE-quantizer residual (TurboQuant_prod) instead of MSE-only quantizer (ablation)TurboQuant · inner-product estimation, DBpedia 1536-dbiasImproved
use TurboQuant_prod instead of TurboQuant_mse at higher bit width (ablation)TurboQuant · inner-product errorerrorMixed
use TurboQuant instead of KIVI/SnapKV/PyramidKV on Needle-in-a-HaystackKV cache compression · Needle-In-A-Haystack recallrecallImproved

Later papers that compared with this method

1 later accepted paper reports TurboQuant as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy