arXiv:2604. 28123v3 Announce Type: replace-cross Abstract: The standard post-training recipe for large multimodal models (LMMs) applies supervised fine-tuning (SFT) on curated demonstrations followed by reinforcement learning with verifiable rewards (RLVR).
VISA (Visual Instruction Synthesis Agent) is an agentic framework that transforms multimodal instruction synthesis into a self‑evolving loop. Each cycle analyzes images to filter constraints, samples new constraint sets, generates candidate instructions, and verifies them using executable tools and large language model judges. Failed samples trigger diagnostic recovery, while accepted samples are evaluated against the target model to estimate difficulty, with all feedback written back to memory to adapt future rounds and provide reward signals for reinforcement learning.
arXiv:2606. 07000v1 Announce Type: new Abstract: Recent post-training methods, particularly Reinforcement Learning with Verifiable Rewards (RLVR), have significantly enhanced the reasoning ability of Large Vision-Language Models (LVLMs).
The Pistis Technical Report introduces the Pistis model family, comprising 27B- and 9B-parameter multimodal large language models built on Qwen3.6 and Qwen3.5. The models are developed through a scalable post‑training framework that begins with large‑scale multimodal supervised fine‑tuning and then applies Interleaved Distillation and Reinforcement Learning (IDRL) to integrate on‑policy distillation and reinforcement learning within a single training loop. Two specialized variants—Pistis‑Thinking for deep multimodal reasoning and Pistis‑Agentic for long‑horizon planning, iterative reasoning, and tool use—are produced at both scales, and a system‑level method called Pistis‑Auto‑Harnessing (PAH) further improves inference harness performance without updating model parameters.
MiMo‑V2.6 is an omni‑modal reinforcement learning framework that scales compute to advance large foundation models toward self‑improvement. It trains on a broad multimodal corpus before RL, then scales RL along three axes: larger batches and throughput, diverse environments across code, general, visual, and cyber domains, and enhanced grader compute for better reward signals. The system incorporates infrastructure such as a frozen MoE router, multi‑layer defense against reward hacking, unified trajectory representation, high‑concurrency rollout, and training‑inference consistency, and it is open‑source to support reproducibility and further research.
arXiv:2609.39168v1 Announce Type: new Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) has improved the reasoning capabilities of Multimodal Large Language Models (MLLMs), yet existing...
arXiv:2608. 08326v1 Announce Type: new Abstract: Reinforcement learning with verifiable rewards (RLVR) has emerged as an effective approach for improving multimodal reasoning.
Open multimodal reasoning models have benefited from large-scale reasoning supervision, yet reliable post-training remains challenging due to uneven data quality, inefficient supervision construction,...
arXiv:2610.23533v1 Announce Type: new Abstract: Recent image generators exhibit rich visual priors that can support dense visual perception. However, existing approaches typically encode multiple per...
VinciCoder is a unified framework for multimodal code generation that addresses the limitations of single-task models by training on a large-scale curated corpus of 1.3 M direct generation pairs and 300 k visual‑refinement tasks. It introduces a coarse‑to‑fine Visual Reinforcement Learning (ViRL) approach that uses visual similarity across multi‑scale patches to provide an implementation‑agnostic reward, improving alignment between rendered outputs and input visuals. Experiments on diverse benchmarks show VinciCoder outperforms existing methods, and ablation studies confirm the effectiveness of ViRL.
arXiv:2512. 03438v3 Announce Type: replace Abstract: Agentic reasoning models trained with multimodal reinforcement learning (MMRL) have become increasingly capable, yet they are almost universally optimized using sparse, outcome-based rewards computed based on the final answers.
The paper introduces a visual-native agent harness that registers every tool‑returned image as an addressable reference, enabling reuse of intermediate visual evidence in multimodal deep search tasks. It also proposes On‑Policy Data Evolution (ODE), a closed‑loop data generator that refines training data across rounds based on policy rollouts, targeting the agent’s evolving needs. Experiments on eight benchmarks show ODE boosts the Qwen3‑VL‑8B agent’s performance from 24.9% to 39.0% on average, and from 30.6% to 41.5% at 30B, while validating the benefits of image bank reuse and rollout‑feedback evolution.