Research

From Generalist to Medical Specialist: Building Domain-Specific Multimodal LLMs via CPT, SFT and RLVR

A systematic overview of a three-stage pipeline for building medical multimodal LLMs from generalist backbones: Continual Pretraining (CPT), Supervised Fine-Tuning (SFT), and Reinforcement Learning with Verifiable Rewards (RLVR). InfiMed2-4B achieves 60.7% average accuracy across medical benchmarks, a 7.0 point improvement over the baseline.

InfiAgent: A self-evolving pyramid agent framework for infinite scenarios released

With the rapid development of artificial intelligence, Large Language Model (LLM) agents have demonstrated remarkable capabilities in organizing and executing complex tasks. However, their development heavily relies on carefully designed workflows, repeatedly debugged prompts, and deep domain expertise. This highly manual approach significantly hinders the large-scale adoption and cost-effectiveness of agent-based technology across industries.

Model Merging Scaling Laws: A New Way to Predict and Plan LLM Composition

We study empirical scaling laws for language model merging measured by cross-entropy. Despite wide practical use, merging has lacked a quantitative rule predicting returns as experts are added or model size scales. We identify a compact power law coupling model size and expert count: a size-dependent loss floor that decreases with capacity and an inverse‑k merging tail with clear diminishing returns. The law holds in-domain and cross-domain, tightly fits curves across architectures and methods (Average, TA, TIES, DARE), and explains two regularities: most gains arrive early and variability shrinks as more experts are included. A simple theory accounts for the ~1/k gain pattern and links floor and tail to base model properties and cross-domain diversity. This enables predictive planning: estimate experts needed for a target loss, decide when to stop, and trade off scaling the base model versus adding experts under a fixed budget—turning merging from heuristic practice into a computationally efficient, plannable alternative to multitask training. It suggests a scaling principle for distributed generative AI: predictable gains via composing specialists, offering a complementary path toward AGI-level systems.

InfiMed: Low-Resource Medical MLLMs with Advancing Understanding and Reasoning

We introduce InfiMed-Series models, InfiMed-SFT-3B and InfiMed-RL-3B, medical-focused Multimodal Large Language Models (MLLMs) developed by the InfiX-AI team. InfiMed-RL-3B achieves an average accuracy of 59.2% across seven authoritative medical benchmarks (including MMMU Health & Medicine, OmniMedVQA, and PMC-VQA), significantly outperforming all comparable-scale models such as MedGemma-4B-IT (54.8%) and even surpassing the larger InternVL3-8B (57.3%).

InfiR2: A Comprehensive FP8 Training Recipe for Reasoning-Enhanced Language Models

The immense computational cost of training Large Language Models (LLMs) presents a major barrier to innovation. While FP8 training offers a promising solution with significant theoretical efficiency gains, its widespread adoption has been hindered by the lack of a comprehensive, open-source training recipe. To bridge this gap, we introduce an end-to-end FP8 training recipe that seamlessly integrates continual pre-training and supervised fine-tuning. Our methodology employs a fine-grained, hybrid-granularity quantization strategy to maintain numerical fidelity while maximizing computational efficiency. Through extensive experiments, including the continue pre-training of models on a 160B-token corpus, we demonstrate that our recipe is not only remarkably stable but also essentially lossless, achieving performance on par with the BF16 baseline across a suite of reasoning benchmarks. Crucially, this is achieved with substantial efficiency improvements, including up to a 22% reduction in training time, a 14% decrease in peak memory usage, and a 19% increase in throughput. Our results establish FP8 as a practical and robust alternative to BF16, and we will release the accompanying code to further democratize large-scale model training.

InfiAlign: A Scalable and Sample-Efficient Framework for Enhancing LLM Reasoning

InfiAlign is a novel framework that combines SFT and DPO with an advanced data selection pipeline to efficiently enhance LLM reasoning capabilities.

InfiGUI-G1: Advancing GUI Grounding with Adaptive Exploration Policy Optimization

We introduce InfiGUI-G1, a multimodal GUI agent that employs Adaptive Exploration Policy Optimization (AEPO) to improve semantic alignment in GUI grounding, achieving up to 8.3% relative improvement over baseline methods.

InfiGFusion: Graph-on-Logits Distillation for Scalable Model Fusion

Recent advances in large language models (LLMs) have intensified efforts to fuse heterogeneous open-source models into a unified system that inherits their complementary strengths. Existing logit-based fusion methods maintain inference efficiency but treat vocabulary dimensions independently, overlooking semantic dependencies encoded by cross-dimension interactions. These dependencies reflect how token types interact under a model's internal reasoning and are essential for aligning models with diverse generation behaviors. To explicitly model these dependencies, we propose InfiGFusion, the first structure-aware fusion framework with a novel Graph-on-Logits Distillation (GLD) loss. Specifically, we retain the top-k logits per output and aggregate their outer products across sequence positions to form a global co-activation graph, where nodes represent vocabulary channels and edges quantify their joint activations. To ensure scalability and efficiency, we design a sorting-based closed-form approximation that reduces the original O(n^4) cost of Gromov-Wasserstein distance to O(n log n), with provable approximation guarantees. Experiments across multiple fusion settings show that GLD consistently improves fusion quality and stability. InfiGFusion outperforms SOTA models and fusion baselines across 11 benchmarks spanning reasoning, coding, and mathematics. It shows particular strength in complex reasoning tasks, with +35.6 improvement on Multistep Arithmetic and +37.06 on Causal Judgement over SFT, demonstrating superior multi-step and relational inference.

InfiFPO: Implicit Model Fusion via Preference Optimization in Large Language Models

Model fusion combines multiple Large Language Models (LLMs) with different strengths into a more powerful, integrated model through lightweight training methods. Existing works on model fusion focus primarily on supervised fine-tuning (SFT), leaving preference alignment (PA) --a critical phase for enhancing LLM performance--largely unexplored. The current few fusion methods on PA phase, like WRPO, simplify the process by utilizing only response outputs from source models while discarding their probability information. To address this limitation, we propose InfiFPO, a preference optimization method for implicit model fusion. InfiFPO replaces the reference model in Direct Preference Optimization (DPO) with a fused source model that synthesizes multi-source probabilities at the sequence level, circumventing complex vocabulary alignment challenges in previous works and meanwhile maintaining the probability information. By introducing probability clipping and max-margin fusion strategies, InfiFPO enables the pivot model to align with human preferences while effectively distilling knowledge from source models. Comprehensive experiments on 11 widely-used benchmarks demonstrate that InfiFPO consistently outperforms existing model fusion and preference optimization methods. When using Phi-4 as the pivot model, InfiFPO improve its average performance from 79.95 to 83.33 on 11 benchmarks, significantly improving its capabilities in mathematics, coding, and reasoning tasks.

InfiGUI-R1: Advancing Multimodal GUI Agents from Reactive Actors to Deliberative Reasoners

InfiGUI-R1 is a multimodal large language model-based GUI agent trained using reinforcement learning to enhance planning and error recovery skills for GUI tasks, achieving state-of-the-art performance on multiple benchmarks.

InfiFusion: A Unified Framework for Enhanced Cross-Model Reasoning via LLM Fusion

InfiGUIAgent: A Multimodal Generalist GUI Agent with Native Reasoning and Reflection

InfiGUIAgent is a multimodal generalist GUI agent trained through a two-stage supervised fine-tuning approach, focusing on fundamental GUI understanding skills and advanced reasoning capabilities for native GUI interactions.