| On the Design of Qwen3.8-Next Architecture: Evaluation, Efficiency, and Training Stability | ✓ Link | 88.5 | Qwen3.8-Flash-Next | 2026-08-31 |
| Qwen3.8-Max: A New Bar for Coding and Cowork | ✓ Link | 88.0 | Qwen3.8-Max | 2026-08-03 |
| Qwen3.8-27B | ✓ Link | 85.9 | Qwen3.8-27B | 2026-08-05 |
| Qwen3.6 | ✓ Link | 85.3 | Qwen3.6-35B-A3B | 2026-04-21 |
| Qwen3.5: Towards Native Multimodal Agents | ✓ Link | 85.1 | Qwen3.5-122B-A10B | 2026-02-16 |
| Qwen3.5: Towards Native Multimodal Agents | ✓ Link | 84.1 | Qwen3.5-35B-A3B | 2026-02-16 |
| Qwen3.5-Omni Technical Report | | 84.1 | Qwen3.5-Omni-Plus | 2026-04-21 |
| Qwen3.6 | ✓ Link | 84.1 | Qwen3.6-27B | 2026-04-21 |
| Qwen3.5: Towards Native Multimodal Agents | ✓ Link | 83.9 | Qwen3.5-397B-A17B | 2026-02-16 |
| Qwen3.5: Towards Native Multimodal Agents | ✓ Link | 83.7 | Qwen3.5-27B | 2026-02-16 |
| Capek 0.5: An Execution-Centric Vision-Language Model for Embodied Intelligence | | 83.4 | Capek0.5-35B-A3B | 2026-08-07 |
| Introducing efficiency focused "Thinking CAP" model series | ✓ Link | 81.9 | ThinkingCap-Qwen3.6-27B | 2026-07-06 |
| HopChain: Multi-Hop Data Synthesis for Generalizable Vision-Language Reasoning | | 81.7 | Qwen3.5-397B-A17B (RLVR w/ Multi-Hop) | 2026-03-17 |
| SpatialASM: Internalizing Active Mentalization for Spatial Reasoning | ✓ Link | 81.7 | SpatialASM-Qwen3.5-9B | 2026-10-09 |
| Qwen3-VL Technical Report | ✓ Link | 81.3 | Qwen3-VL-235B-A22B-Thinking | 2025-11-26 |
| Qwen3.5: Towards Native Multimodal Agents | ✓ Link | 80.3 | Qwen3.5-9B | 2026-02-16 |
| Cosmos 3: Omnimodal World Models for Physical AI | ✓ Link | 79.2 | Cosmos3-Super | 2026-06-01 |
| Qwen-Drive-1.0: An Initial Step towards a Vision-Language Foundation Model for Autonomous Driving | ✓ Link | 78.95 | Qwen-Drive-1.0-SFT | 2026-08-31 |
| InternVL3: Exploring Advanced Training and Test-Time Recipes for Open-Source Multimodal Models | ✓ Link | 78.0 | InternVL3-78B | 2025-04-14 |
| Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution | ✓ Link | 77.8 | Qwen2-VL-72B | 2024-09-18 |
| Molmo2: Open Weights and Data for Vision-Language Models with Video Understanding and Grounding | ✓ Link | 77.6 | Molmo2-8B | 2026-01-15 |
| iVGR: Internalizing Visually Grounded Reasoning for MLLMs with Reinforcement Learning | ✓ Link | 76.3 | iVGR-Qwen3-VL-32B (ours) | 2026-05-29 |
| LongCat-Flash-Omni Technical Report | ✓ Link | 76.0 | LongCat-Flash-Omni | 2025-10-31 |
| SenseNova-U1.5: Towards Native Unified Visual Intelligence | ✓ Link | 75.8 | SenseNova-U1.5-8B-MoT | 2026-09-10 |
| Qwen2.5-VL Technical Report | ✓ Link | 75.7 | Qwen2.5-VL-72B | 2025-02-19 |
| STEP3-VL-10B Technical Report | ✓ Link | 75.56 | Step3-VL-10B (PaCoRe) | 2026-01-14 |
| Molmo and PixMo: Open Weights and Open Data for State-of-the-Art Vision-Language Models | ✓ Link | 75.2 | Molmo 72B | 2024-09-25 |
| PerceptionLM: Open-Access Data and Models for Detailed Visual Understanding | ✓ Link | 75.0 | PLM-8B | 2025-04-17 |
| TUNA: Taming Unified Visual Representations for Native Unified Multimodal Models | ✓ Link | 74.7 | Tuna | 2025-12-01 |
| Youtu-VL: Unleashing Visual Potential via Unified Vision-Language Supervision | ✓ Link | 74.6 | Youtu-VL (4B) | 2026-01-27 |
| From Seeing to Thinking: Decoupling Perception and Reasoning Improves Post-Training of Vision-Language Models | ✓ Link | 74.5 | Qwen3-VL-8B (staged training) | 2026-05-19 |
| Awaking Spatial Intelligence in Unified Multimodal Understanding and Generation | ✓ Link | 74.0 | JoyAI-Image-Und | 2026-05-05 |
| PerceptionDLM: Parallel Region Perception with Multimodal Diffusion Language Models | ✓ Link | 73.7 | PerceptionDLM-Base (8B) | 2026-06-17 |
| Introducing Command A Vision: Multimodal AI built for business | | 73.6 | Command A Vision | 2025-07-31 |
| Kwai Keye-VL 1.5 Technical Report | ✓ Link | 73.5 | Keye-VL-1.5-8B | 2025-09-01 |
| vLLM Semantic Router: Signal Driven Decision Routing for Mixture-of-Modality Models | ✓ Link | 73.4 | d3 | 2026-02-23 |
| Vero: An Open RL Recipe for General Visual Reasoning | ✓ Link | 73.3 | Vero-Qwen3I-8B | 2026-04-06 |
| ERNIE 4.5 Technical Report | ✓ Link | 73.2 | ERNIE-4.5-VL-424B-A47B | 2025-06-30 |
| SwimBird: Eliciting Switchable Reasoning Mode in Hybrid Autoregressive MLLMs | ✓ Link | 73.1 | SwimBird | 2026-02-05 |
| Robust-U1: Can MLLMs Self-Recover Corrupted Visual Content for Robust Understanding? | ✓ Link | 72.81 | Robust-U1 | 2026-06-06 |
| VideoLLaMA 3: Frontier Multimodal Foundation Models for Image and Video Understanding | ✓ Link | 72.7 | VideoLLaMA3 7B | 2025-01-22 |
| MiniCPM-V 4.5: Cooking Efficient MLLMs via Architecture, Data, and Training Recipe | ✓ Link | 72.1 | MiniCPM-V 4.5 | 2025-09-16 |
| LongCat-Next: Lexicalizing Modalities as Discrete Tokens | ✓ Link | 72.0 | LongCat-Next (A3B, 68.5B) | 2026-03-29 |
| LLaVA-OneVision: Easy Visual Task Transfer | ✓ Link | 71.9 | LLaVA-OneVision-72B | 2024-08-06 |
| Learning from Primitive: Probing Visual Reasoning of LVLMs via Counting | | 71.83 | Qwen3-VL-8B (Learning from Primitive) (extra data) | 2026-09-08 |
| DeepAlign: Mitigating Modality Conflict through Modality-Specific Alignment | | 71.6 | DeepAlign (InternVL3-8B) | 2026-06-03 |
| Unfold The World: Factorize 4D Properties in Reinforcing Spatial Reasoning | ✓ Link | 71.6 | FactoSR-8B-RL | 2026-09-03 |
| Innovator-VL: A Multimodal Large Language Model for Scientific Discovery | ✓ Link | 71.5 | Innovator-VL (8B-Instruct) | 2026-01-27 |
| Penguin-VL: Exploring the Efficiency Limits of VLM with LLM-based Vision Encoders | ✓ Link | 71.5 | Penguin-VL 8B | 2026-03-06 |
| LFM2 Technical Report | ✓ Link | 71.37 | LFM2-VL-3B | 2025-11-28 |
| Thinking in Uncertainty: Mitigating Hallucinations in MLRMs with Latent Entropy-Aware Decoding | ✓ Link | 71.2 | VL-Rethinker-7B with LEAD | 2026-03-09 |
| Odysseus: Scaling VLMs to 100+ Turn Decision-Making in Games via Reinforcement Learning | | 71.11 | Odysseus (RL on SFT) | 2026-05-01 |
| ClaimDiff-RL: Fine-Grained Caption Reinforcement Learning through Visual Claim Comparison | ✓ Link | 71.1 | ClaimDiff-RL (actor-only) (init: Qwen3VL-32B-Instruct) | 2026-05-19 |
| OpenSpatial: A Principled Data Engine for Empowering Spatial Intelligence | ✓ Link | 71.0 | OpenSpatial-Qwen3-VL-8B | 2026-04-08 |
| LoopVL: Recurrent Visual Intelligence | ✓ Link | 70.98 | LoopVL-1B | 2026-09-29 |
| Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context | ✓ Link | 70.4 | Gemini 1.5 Pro | 2024-03-08 |
| Qwen2.5-Omni Technical Report | ✓ Link | 70.3 | Qwen2.5-Omni-7B | 2025-03-26 |
| Generation as Auxiliary Supervision: Enhancing Visual Understanding at Zero Inference Overhead via Decoupled Embedding Prediction | | 70.3 | GAS (4B) | 2026-08-12 |
| NaVLM-PVC: Progressive Visual Compression for Efficient Native-Resolution Encoding in MLLMs | | 70.3 | NaVLM-PVC | 2026-09-08 |
| Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution | ✓ Link | 70.1 | Qwen2-VL-7B | 2024-09-18 |
| Kimi-VL Technical Report | ✓ Link | 70.0 | Kimi-VL-A3B-Thinking-2506 | 2025-04-10 |
| LLaVA-OneVision-2: Towards Next-Generation Perceptual Intelligence | ✓ Link | 69.7 | LLaVA-OneVision-2-8B | 2026-05-25 |
| BLIP3-o: A Family of Fully Open Unified Multimodal Models-Architecture, Training and Dataset | ✓ Link | 69.0 | BLIP3-o (8B) | 2025-05-14 |
| Grok-1.5 Vision Preview | | 68.7 | Grok-1.5V | 2024-04-12 |
| Qwen2.5-VL Technical Report | ✓ Link | 68.5 | Qwen2.5-VL-7B | 2025-02-19 |
| Look Where It Matters: High-Resolution Crops Retrieval for Efficient VLMs | ✓ Link | 68.5 | AwaRes | 2026-03-14 |
| ROSE: Rotate Your Large Language Model to See | | 68.2 | ROSE-7B (extra data) | 2026-06-03 |
| Kimi-VL Technical Report | ✓ Link | 68.1 | Kimi-VL-A3B-Instruct | 2025-04-10 |
| DiffusionVL: Translating Any Autoregressive Models into Diffusion Vision Language Models | ✓ Link | 68.0 | DiffusionVL | 2025-12-17 |
| MOSS-VL Technical Report | ✓ Link | 68.0 | MOSS-VL-Instruct-0708 | 2026-08-15 |
| From Pixels to Words -- Towards Native One-Vision Models at Scale | ✓ Link | 67.8 | NEO-ov 9B | 2026-05-27 |
| Robust-R1: Degradation-Aware Reasoning for Robust Visual Understanding | ✓ Link | 67.71 | Robust-R1 (SFT and RL) | 2025-12-19 |
| Starve to Perceive: Taming Lazy Perception in VLMs with Constrained Visual Bandwidth | | 67.71 | Starve to Perceive | 2026-05-18 |
| Kwai Keye-VL Technical Report | ✓ Link | 67.7 | Keye-VL-8B | 2025-07-02 |
| Faithful GRPO: Improving Visual Spatial Reasoning in Multimodal Language Models via Constrained Policy Optimization | | 67.64 | FGRPO (7B) | 2026-04-09 |
| VisionFoundry: Teaching VLMs Visual Perception with Synthetic Images | ✓ Link | 67.5 | MiMo-VL-7B-SFT + VisionFoundry-10K | 2026-04-10 |
| AdaptVision: Efficient Vision-Language Models via Adaptive Visual Acquisition | ✓ Link | 67.32 | AdaptVision (Qwen2.5-VL-7B-Instruct) | 2025-12-03 |
| VideoLLaMA 3: Frontier Multimodal Foundation Models for Image and Video Understanding | ✓ Link | 67.3 | VideoLLaMA3 2B | 2025-01-22 |
| InfiniteVL: Synergizing Linear and Sparse Attention for Highly-Efficient, Unlimited-Input Vision-Language Models | ✓ Link | 67.3 | InfiniteVL-Base | 2025-12-09 |
| FastVLM: Efficient Vision Encoding for Vision Language Models | ✓ Link | 67.2 | FastVLM-7B | 2024-12-17 |
| SpatialEvo: Self-Evolving Spatial Intelligence via Deterministic Geometric Environments | ✓ Link | 66.7 | SpatialEvo (Qwen2.5-VL-7B-Instruct) | 2026-04-15 |
| LLaDA2.0-Uni: Unifying Multimodal Understanding and Generation with Diffusion Large Language Model | ✓ Link | 66.7 | LLaDA2.0-Uni (16B) | 2026-04-22 |
| Representation Forcing for Bottleneck-Free Unified Multimodal Models | | 66.6 | VAE+RF | 2026-05-29 |
| Dream-VL & Dream-VLA: Open Vision-Language and Vision-Language-Action Models with Diffusion Language Model Backbone | ✓ Link | 66.3 | Dream-VL | 2025-12-27 |
| Generation Enhances Understanding in Unified Multimodal Models via Multi-Representation Generation | ✓ Link | 66.01 | OpenUni-3.6B + UniMRG | 2026-01-29 |
| ZAYA1-VL-8B Technical Report | ✓ Link | 65.0 | ZAYA1-VL-8B | 2026-05-08 |
| Introducing LFM2.5: The Next Generation of On-Device AI | | 64.84 | LFM2.5-VL-1.6B | 2026-01-05 |
| Penguin-VL: Exploring the Efficiency Limits of VLM with LLM-based Vision Encoders | ✓ Link | 63.9 | Penguin-VL 2B | 2026-03-06 |
| CASA: Cross-Attention via Self-Attention for Efficient Vision-Language Fusion | ✓ Link | 63.8 | CAt+v (Qwen2.5-VL, 3B) | 2025-12-22 |
| MiniCPM-V: A GPT-4V Level MLLM on Your Phone | ✓ Link | 63.5 | MiniCPM-Llama3-V 2.5 | 2024-08-03 |
| LLaDA-UI: Bringing Block-wise Diffusion to Vision-Language GUI Agents | ✓ Link | 63.5 | LLaDA-UI (pre-training) | 2026-09-09 |
| LLaDA-V: Large Language Diffusion Models with Visual Instruction Tuning | ✓ Link | 63.2 | LLaDA-V | 2025-05-22 |
| Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution | ✓ Link | 62.9 | Qwen2-VL-2B | 2024-09-18 |
| Intern Lumina U2: A Multi-Codebook Diffusion Large Language Model for Omni-Visual Understanding and Image Generation | ✓ Link | 62.9 | Intern Lumina U2 (Ascend) | 2026-10-14 |
| Ovis: Structural Embedding Alignment for Multimodal Large Language Model | ✓ Link | 62.7 | Ovis-Qwen1.5-14B | 2024-05-31 |
| Falcon2-11B Technical Report | | 61.7 | Falcon2-11B VLM | 2024-07-20 |
| Differences That Matter: Auditing Models for Capability Gap Discovery and Rectification | | 61.4 | Gemma3 + AuditDM | 2025-12-18 |
| CLEAR: Unlocking Generative Potential for Degraded Image Understanding in Unified Multimodal Models | ✓ Link | 61.05 | CLEAR-RL (Bagel-7B) | 2026-04-06 |
| Cheers: Decoupling Patch Details from Semantic Representations Enables Unified Multimodal Comprehension and Generation | ✓ Link | 60.9 | Cheers (1.5B) | 2026-03-13 |
| Qwen-3D: A Generalist 3D Vision-Language Model for Spatial Understanding | ✓ Link | 60.4 | Qwen-3D-3B | 2026-08-04 |
| GeMoE: Gating Entropy is All You Need for Uncertainty-aware Adaptive Routing in MoE-based Large Vision-Language Models | ✓ Link | 60.1 | GeMoE (extra data) | 2026-06-24 |
| Preventing Expert Collapse in MoE-dVLMs via Modality-Wise Norm Alignment | | 58.9 | TowerAlign | 2026-09-08 |
| LFM2.5-VL-450M: Structured Visual Intelligence, Edge to Cloud | | 58.43 | LFM2.5-VL-450M | 2026-04-08 |
| MIRROR: Aligning Semantic Relations from Language to Image via Gromov--Wasserstein | | 57.7 | LLaVA-NeXT-7B + MIRROR | 2026-06-28 |
| Emu3: Next-Token Prediction is All You Need | ✓ Link | 57.4 | Emu3 | 2024-09-27 |
| NeuPAT: Neuron-aware Plasticity Allocation Tuning for Language-Preserving MLLMs | | 56.99 | NeuPAT (Qwen3-4B-Instruct-2507) (extra data) | 2026-08-08 |
| Rethinking Projector Training in Multimodal LLMs | ✓ Link | 56.9 | PORTAL (LLaMA 3.1 8B + CLIP) | 2026-10-13 |
| VISion On Request: Enhanced VLLM efficiency with sparse, dynamically selected, vision-language interactions | | 54.6 | VISOR | 2026-03-24 |
| ESC: Emotional Self-Correction for Reliable Vision-Language Models | ✓ Link | 54.51 | ESC + Qwen2-VL-7B | 2026-07-01 |
| SpatialCLIP: Learning 3D-aware Image Representations from Spatially Discriminative Language | | 54.2 | Spatial-LLaVA-7B (extra data) | 2025-06-11 |
| LinguDistill: Recovering Linguistic Ability in Vision- Language Models via Selective Cross-Modal Distillation | | 49.0 | LinguDistill | 2026-04-01 |
| Modality Gap-Driven Subspace Alignment Training Paradigm For Multimodal Large Language Models | ✓ Link | 47.97 | ReVision | 2026-02-02 |
| Unfold The World: Factorize 4D Properties in Reinforcing Spatial Understanding | | 47.3 | FactoSR-8B-RL | 2026-09-08 |
| T5Gemma 2: Seeing, Reading, and Understanding Longer | | 46.1 | T5Gemma 2 4B-4B | 2025-12-16 |