Multimodal large language models (MLLMs) have shown increasing potential for applications in the financial domain. However, the challenges posed by real-world financial scenarios, including high-information-density, strong noise interference, and cross-modal multi-hop reasoning, go beyond the evaluation scope of existing financial multimodal benchmarks. To this end, we propose UniFinEval, the first unified multimodal financial evaluation benchmark designed for high-density financial information environments, covering three modalities: text, image, and video. UniFinEval abstracts three progressive core scenarios from perception to reasoning and decision-making, and further refines them into six task categories. It contains 5,567 high-quality Chinese and English question–answer (QA) pairs constructed by financial experts. We systematically evaluate 10 mainstream MLLMs under zero-shot and chain-of-thought (CoT) settings. Experimental results show that Gemini-3-Pro achieves the best overall performance, but still exhibits a significant gap compared with financial experts. Further error analysis reveals the capability deficiencies of MLLMs in financial scenarios. UniFinEval aims to systematically evaluate the capabilities of MLLMs in real financial scenarios, thereby promoting the application of MLLMs in the financial domain. The related code and data are available at https://github.com/aifinlab/UniFinEval.

High-quality mathematical and logical datasets with verifiable answers are essential for strengthening the reasoning capabilities of large language models (LLMs). While recent data augmentation techniques have facilitated the creation of large-scale benchmarks, existing LLM-generated datasets often suffer from limited reliability, diversity, and scalability. To address these challenges, we introduce PuzzleClone, a formal framework for synthesizing verifiable data at scale using a novel DSL-driven approach. Our approach features three key innovations: (1) encoding seed puzzles into structured logical specifications, (2) generating scalable variants through systematic variable and constraint randomization, and (3) ensuring validity via a reproduction mechanism. Applying PuzzleClone, we construct PC-83K, a benchmark comprising over 83K diverse and programmatically validated puzzles. The generated puzzles span a wide spectrum of difficulty and formats, posing significant challenges to current state-of-the-art models. Experimental results show that post training (SFT and RL) on PC-83K yields substantial improvements not only on the testset but also on various logic and mathematical benchmarks. Post training raises average performance on PC-83K from 14.5 to 66.0 and delivers consistent improvements across 7 logic and mathematical benchmarks up to 18.4 absolute percentage points (SATBench from 51.6 to 70.0). Our code and data are available at https://github.com/HiThink-Research/PuzzleClone.

Reasoning capability plays a significantly critical role in the the broad applications of Large Language Models (LLMs). To enhance the reasoning performance of LLMs, diverse Reinforcement Learning (RL)-based fine-tuning approaches have been proposed to address the limited generalization capability of LLMs trained solely via Supervised Fine-Tuning (SFT). Despite their effectiveness, two major limitations hinder the advancement of LLMs. First, vanilla RL-based approaches ignore annotated Chain-of-Thought (CoT) and incorporate unstable reasoning path sampling, which typically results in model collapse, unstable training process, and suboptimal performance. Second, existing SFT approaches generally overemphasize the annotated CoT, potentially leading to performance degradation due to insufficient exploitation of potential CoT. In this paper, we propose a Contrastive learning with annotated CoT-based Reinforced Fine-Tuning approach, i.e., CARFT, to enhance the reasoning performance of LLMs while addressing the aforementioned limitations. Specifically, we propose learning a representation for each CoT. Based on this representation, we design novel contrastive signals to guide the fine-tuning process. Our approach not only fully exploits the available annotated CoT but also stabilizes the fine-tuning procedure by incorporating an additional unsupervised learning signal. We conduct comprehensive experiments and in-depth analysis with three baseline approaches, two foundation models, and two datasets to demonstrate significant advantages of CARFT in terms of robustness, performance (up to 10.15%), and efficiency (up to 30.62%).