AI Paper Atlas

Research map / Architectures and efficiency / Quantization

AnyBCQ: Hardware Efficient Flexible Binary-Coded Quantization for Multi-Precision LLMs

Gunho Park, Jeongin Bae, Beomseok Kwon, Byeongwook Kim, Se Jung Kwon, Dongsoo Lee

ICLR 2026 · Architectures and efficiency › Quantization · approach: Post-training quantization of LLMs · 4 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
use fixed-precision optimization instead of progressive multi-precision at 3 bits (ablation)AnyBCQ · Wikitext perplexity (3-bit, Llama-3.1-8B)perplexity: 8.08 -> 7.68Hurt
use fixed-precision optimization instead of progressive multi-precision at 4 bits (ablation)AnyBCQ · Wikitext perplexity (4-bit, Llama-3.1-8B)perplexity: 6.84 -> 6.62Hurt
use fixed-precision optimization instead of multi-precision at 2 bits (ablation)AnyBCQ · Wikitext perplexity (2-bit, Llama-3.1-8B)perplexity: 19.01 -> 18.95No clear effect
share binary bit-planes across precisions instead of separate modelsAnyBCQ · Llama-3.1-8B memory footprintmemory (GB): 9.85 -> 4.99Improved
replace Any-Precision LLM with AnyBCQ at 2 bitsLlama-3.1-8B · MMLU 5-shot (2-bit)accuracy: 24.66 -> 35.32Improved
replace Any-Precision LLM with AnyBCQ at 2 bitsLlama-3.1-8B · Wikitext perplexity (2-bit)perplexity: 1680.77 -> 19.01Improved
replace Any-Precision LLM with AnyBCQ at 4 bitsLlama-3.1-8B · MMLU 5-shot (4-bit)accuracy: 64.04 -> 63.15Hurt
replace Any-Precision LLM with AnyBCQ at 4 bitsLlama-3.1-8B · Wikitext perplexity (4-bit)perplexity: 6.70 -> 6.84Hurt
replace Any-Precision LLM kernel with AnyBCQ kernel (2-bit)Llama-3.1-8B · decoding throughput A100tokens/s: 228 -> 245Improved
replace Any-Precision LLM kernel with AnyBCQ kernel (N=28672, K=8192, 2-bit)GEMV kernel · GEMV latencylatency (us): 830 -> 747Improved
replace Any-Precision LLM with AnyBCQ in mixed-precision decoding at 3.6 bitsPMPD setting · CNN/DailyMailROUGE-L: 0.154 -> 0.178Improved
replace Any-Precision LLM with AnyBCQ in mixed-precision decoding at 2.23 bitsPMPD setting · CNN/DailyMailROUGE-L: 0.097 -> 0.113Improved

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