Beyond Gemini-3-Pro: Revisiting LLM Routing and Aggregation at Scale

• URL: http://arxiv.org/abs/2601.01330v1
• Date: Sun, 04 Jan 2026 02:05:52 GMT
• Title: Beyond Gemini-3-Pro: Revisiting LLM Routing and Aggregation at Scale
• Abstract summary: Large Language Models (LLMs) have rapidly advanced, with Gemini-3-Pro setting a new performance milestone.<n>We explore collective intelligence as an alternative to monolithic scaling, and demonstrate that open-source LLMs' collaboration can surpass Gemini-3-Pro.
• Score: 29.165198842583575
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Large Language Models (LLMs) have rapidly advanced, with Gemini-3-Pro setting a new performance milestone. In this work, we explore collective intelligence as an alternative to monolithic scaling, and demonstrate that open-source LLMs' collaboration can surpass Gemini-3-Pro. We first revisit LLM routing and aggregation at scale and identify three key bottlenecks: (1) current train-free routers are limited by a query-based paradigm focusing solely on textual similarity; (2) recent aggregation methods remain largely static, failing to select appropriate aggregators for different tasks;(3) the complementarity of routing and aggregation remains underutilized. To address these problems, we introduce JiSi, a novel framework designed to release the full potential of LLMs' collaboration through three innovations: (1) Query-Response Mixed Routing capturing both semantic information and problem difficulty; (2) Support-Set-based Aggregator Selection jointly evaluating the aggregation and domain capacity of aggregators; (3) Adaptive Routing-Aggregation Switch dynamically leveraging the advantages of routing and aggregation. Comprehensive experiments on nine benchmarks demonstrate that JiSi can surpass Gemini-3-Pro with only 47% costs by orchestrating ten open-source LLMs, while outperforming mainstream baselines. It suggests that collective intelligence represents a novel path towards Artificial General Intelligence (AGI).

Related papers

• Empowering Compact LLMs with Fusion of Layer-wise Exits for Recommendation [59.216721553953]
  Large language model-based recommender systems (LLM-RSs) have demonstrated remarkable capabilities, but are computationally unsustainable for many real-world applications.<n>We propose a discriminative framework that enhances compact LLMs while retaining scalable full-corpus ranking.<n>Experiments on three real-world datasets with Qwen 3 1.7B and Llama 3.2 3B show that FLEXRec achieves state-of-the-art accuracy while remaining highly efficient.
  arXiv  Detail & Related papers  (2026-08-18T03:14:28Z)
• Harnessing the Wisdom of LLM Crowds through Complementarity-Driven Iterative Collaboration [7.252833168419753]
  Large language models (LLMs) are increasingly deployed in enterprise settings, yet individual models remain bounded by model-specific capability limitations.<n>We propose WILC (Wisdom Integration of LLM Crowds), a framework grounded in two design principles.<n>WilC matches the average benchmark performance of GPT-5.2 at roughly 7 times lower estimated per-Query cost.
  arXiv  Detail & Related papers  (2026-07-31T07:13:59Z)
• Mixture of Complementary Agents for Robust LLM Ensemble [28.210955621253863]
  Multi-AI collaboration, such as ensembling or debating large language models (LLMs), is a promising paradigm for aggregating information and boosting performance.<n>Existing approaches primarily focus on accuracy (picking the strongest models) or diversity (ensuring variety)<n>We reframe proposer selection as a selection problem akin to feature selection, where the value of an LLM lies in its complementarity with others.
  arXiv  Detail & Related papers  (2026-05-21T21:29:05Z)
• UniCVR: From Alignment to Reranking for Unified Zero-Shot Composed Visual Retrieval [87.12447641908143]
  We propose UniCVR, the first unified zero-shot composed visual retrieval framework.<n>UniCVR jointly addresses all three tasks without any task-specific human-annotated data.<n>In Stage I, we train the MLLM as a compositional query embedder via contrastive learning on a curated dataset of approximately 3.5M samples.<n>In Stage II, we introduce an MLLM-guided dual-level reranking mechanism that applies adaptive budgeted subset scoring to a small number of top-ranked candidates.
  arXiv  Detail & Related papers  (2026-04-22T08:16:50Z)
• TRINITY: An Evolved LLM Coordinator [20.55517425459279]
  Trinity is a lightweight coordinator that orchestrates collaboration among large language models (LLMs)<n>It processes queries over multiple turns, where at each turn the coordinator assigns one of three roles to a selected LLM.<n>Experiments show that Trinity consistently outperforms individual models and existing methods across coding, math, reasoning, and domain knowledge tasks.
  arXiv  Detail & Related papers  (2025-12-04T11:45:21Z)
• ClusterFusion: Hybrid Clustering with Embedding Guidance and LLM Adaptation [52.794544682493814]
  Large language models (LLMs) provide strong contextual reasoning, yet prior work mainly uses them as auxiliary modules to refine embeddings or adjust cluster boundaries.<n>We propose ClusterFusion, a hybrid framework that treats the LLM as the clustering core, guided by lightweight embedding methods.<n> Experiments on three public benchmarks and two new domain-specific datasets demonstrate that ClusterFusion achieves state-of-the-art performance on standard tasks.
  arXiv  Detail & Related papers  (2025-12-04T00:49:43Z)
• Run, Ruminate, and Regulate: A Dual-process Thinking System for Vision-and-Language Navigation [52.11339614452127]
  Vision-and-Language Navigation (VLN) requires an agent to dynamically explore complex 3D environments following human instructions.<n>Recent research underscores the potential of harnessing large language models (LLMs) for VLN, given their commonsense knowledge and general reasoning capabilities.<n>We propose a novel dual-process thinking framework dubbed R3, integrating LLMs' generalization capabilities with VLN-specific expertise in a zero-shot manner.
  arXiv  Detail & Related papers  (2025-11-18T04:32:00Z)
• MARAG-R1: Beyond Single Retriever via Reinforcement-Learned Multi-Tool Agentic Retrieval [50.30107119622642]
  Large Language Models (LLMs) excel at reasoning and generation but are inherently limited by static pretraining data.<n>Retrieval-Augmented Generation (RAG) addresses this issue by grounding LLMs in external knowledge.<n>MarAG-R1 is a reinforcement-learned multi-tool RAG framework that enables LLMs to dynamically coordinate multiple retrieval mechanisms.
  arXiv  Detail & Related papers  (2025-10-31T15:51:39Z)
• RCR-Router: Efficient Role-Aware Context Routing for Multi-Agent LLM Systems with Structured Memory [57.449129198822476]
  RCR is a role-aware context routing framework for multi-agent large language model (LLM) systems.<n>It dynamically selects semantically relevant memory subsets for each agent based on its role and task stage.<n>A lightweight scoring policy guides memory selection, and agent outputs are integrated into a shared memory store.
  arXiv  Detail & Related papers  (2025-08-06T21:59:34Z)
• FusionFactory: Fusing LLM Capabilities with Multi-LLM Log Data [60.09659670497899]
  Large language models (LLMs) have a diverse landscape of models, each excelling at different tasks.<n>This diversity drives researchers to employ multiple LLMs in practice, leaving behind valuable multi-LLM log data.<n>We argue that practical fusion must meet two essential requirements: (1) compatibility with real-world serving scenarios (e.g., local and API-based serving), and (2) flexibility to operate at different stages of the LLM pipeline to meet varied user needs.
  arXiv  Detail & Related papers  (2025-07-14T17:58:02Z)
• Agent-UniRAG: A Trainable Open-Source LLM Agent Framework for Unified Retrieval-Augmented Generation Systems [4.683612295430957]
  This paper presents a novel approach for unified retrieval-augmented generation (RAG) systems using the recent emerging large language model (LLM) agent concept.<n>We propose a trainable agent framework called Agent-UniRAG for unified retrieval-augmented LLM systems.<n>The main idea is to design an LLM agent framework to solve RAG tasks step-by-step based on the complexity of the inputs.
  arXiv  Detail & Related papers  (2025-05-28T16:46:31Z)
• C-3PO: Compact Plug-and-Play Proxy Optimization to Achieve Human-like Retrieval-Augmented Generation [13.120930059424975]
  C-3PO is a proxy-centric framework that facilitates communication between retrievers and large language models.<n>Our framework implements three specialized agents that collaboratively optimize the entire RAG pipeline.
  arXiv  Detail & Related papers  (2025-02-10T07:04:32Z)
• Federated Fine-Tuning of LLMs: Framework Comparison and Research Directions [59.5243730853157]
  Federated learning (FL) provides a privacy-preserving solution for fine-tuning pre-trained large language models (LLMs) using distributed private datasets.<n>This article conducts a comparative analysis of three advanced federated LLM (FedLLM) frameworks that integrate knowledge distillation (KD) and split learning (SL) to mitigate these issues.
  arXiv  Detail & Related papers  (2025-01-08T11:37:06Z)
• Improving Generated and Retrieved Knowledge Combination Through Zero-shot Generation [41.43397783169612]
  Open-domain Question Answering (QA) has garnered substantial interest by combining faithfully retrieved passages and relevant passages generated through Large Language Models (LLMs)<n>There is a lack of definitive labels available to pair these sources of knowledge.<n>We propose Bi-Reranking for Merging Generated and Retrieved Knowledge (BRMGR), which utilizes re-ranking methods for both retrieved passages and LLM-generated passages.
  arXiv  Detail & Related papers  (2024-12-25T06:40:36Z)
• Learning to Use Tools via Cooperative and Interactive Agents [58.77710337157665]
  Tool learning empowers large language models (LLMs) as agents to use external tools and extend their utility. We propose ConAgents, a Cooperative and interactive Agents framework, which coordinates three specialized agents for tool selection, tool execution, and action calibration separately. Our experiments on three datasets show that the LLMs, when equipped with ConAgents, outperform baselines with substantial improvement.
  arXiv  Detail & Related papers  (2024-03-05T15:08:16Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.