AI Paper Atlas

Research map / Large language models / Agents, code and math

SLM-MUX: Orchestrating Small Language Models for Reasoning

Chenyu Wang, Zishen Wan, Hao Kang, Emma Chen, Zhiqiang Xie, Tushar Krishna, Vijay Janapa Reddi, Yilun Du

ICLR 2026 · Large language models › Agents, code and math · approach: Mathematical and structural reasoning benchmarks · 11 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 SLM-MUX (confidence-based selection, no discussion) instead of discussion-based orchestrationSLM orchestration · MATH (Llama 3.1 8B, Mistral 8x7B, Gemma 2 27B)accuracy (%): 51.6 -> 61.8 (vs LLM-Debate)Improved
use SLM-MUX instead of Mixture-of-Agents / Debate / VerificationSLM orchestration · GPQAaccuracy (%): 36.8 -> 42.1 (vs LLM-Debate)Improved
discussion-based orchestration (MoA, Debate, Verification) with SLMsLlama 3.1 8B, Mistral 8x7B, Gemma 2 27B · MATHaccuracy (%): best single 56.8 -> 48.4 to 51.6Hurt
use SLM-MUX instead of self-consistency of best single modelSLM-MUX · GPQAaccuracy (%): 42.4 -> 42.1No clear effect
model selection search for complementary pairs (ablation)SLM-MUX · MATHaccuracy (%): 75.5 -> 80.0Improved
use SLM-MUX over Agent Forest (2 samples)SLM-MUX · GPQAaccuracy (%): 40.4 -> 46.3Improved
use SLM-MUX over Agent Forest at best sample budgetSLM-MUX · MATHaccuracy (%): 79.2 -> 79.5No clear effect
draw more samples per model (2 to 9) (ablation)SLM-MUX · MATH, GPQA, GSM8Kaccuracy (%)Improved
add more participating models beyond 2 (ablation)SLM-MUX · GPQAaccuracy (%): GPQA test: 47.98 (K=1), 50.21 (K=2), 51.81 (K=3), 45.18 (K=5)Mixed
use random model selection instead of search (ablation)SLM-MUX · MATH-500 (K=2)accuracy (%): 66.1 -> 67.2Improved
tune the Mixture-of-Agents aggregator prompt (ablation)Mixture-of-Agents · MATHaccuracy (%): 46.2 -> 48.4 (best tuned)No clear effect
apply SLM-MUX to frontier LLMs (GPT-4o + Gemini-2.5-Flash)SLM-MUX · GPQAaccuracy (%): 51.1 -> 60.1Improved

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