| STEP3-VL-10B Technical Report | ✓ Link | 92.38 | Step3-VL-10B (PaCoRe) | 2026-01-14 |
| Kwai Keye-VL 1.5 Technical Report | ✓ Link | 92.0 | Keye-VL-1.5-8B | 2025-09-01 |
| SenseNova-U1: Unifying Multimodal Understanding and Generation with NEO-unify Architecture | ✓ Link | 91.59 | SenseNova-U1-A3B-MoT | 2026-05-12 |
| HopChain: Multi-Hop Data Synthesis for Generalizable Vision-Language Reasoning | | 91.56 | Qwen3.5-397B-A17B (RLVR w/ Multi-Hop) | 2026-03-17 |
| SenseNova-U1.5: Towards Native Unified Visual Intelligence | ✓ Link | 90.47 | SenseNova-U1.5 8B-Think | 2026-09-10 |
| LongCat-Flash-Omni Technical Report | ✓ Link | 89.8 | LongCat-Flash-Omni | 2025-10-31 |
| Step-GUI Technical Report | ✓ Link | 89.1 | Step-GUI-8B | 2025-12-17 |
| InternVL3: Exploring Advanced Training and Test-Time Recipes for Open-Source Multimodal Models | ✓ Link | 89.0 | InternVL3-78B | 2025-04-14 |
| MMFineReason: Closing the Multimodal Reasoning Gap via Open Data-Centric Methods | ✓ Link | 88.9 | MMFineReason (8B) | 2026-01-29 |
| Qwen2.5-VL Technical Report | ✓ Link | 88.6 | Qwen2.5-VL-72B | 2025-02-19 |
| Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution | ✓ Link | 86.5 | Qwen2-VL-72B | 2024-09-18 |
| LLaVA-UHD v4: What Makes Efficient Visual Encoding in MLLMs? | ✓ Link | 86.2 | LLaVA-UHD v4 | 2026-05-09 |
| LLaVA-OneVision-2: Towards Next-Generation Perceptual Intelligence | ✓ Link | 85.7 | LLaVA-OneVision-2-8B | 2026-05-25 |
| Unfold The World: Factorize 4D Properties in Reinforcing Spatial Reasoning | ✓ Link | 85.7 | FactoSR-8B-RL | 2026-09-03 |
| VQRAE: Representation Quantization Autoencoders for Multimodal Understanding, Generation and Reconstruction | | 85.1 | VQRAE (InternViT-300M, Qwen2.5-7B, 448px) | 2025-11-28 |
| From Pixels to Words -- Towards Native One-Vision Models at Scale | ✓ Link | 85.1 | NEO-ov 9B | 2026-05-27 |
| Emerging Properties in Unified Multimodal Pretraining | ✓ Link | 85.0 | BAGEL (7B MoT) | 2025-05-20 |
| Language-guided Frequency Modulation for Large Vision-Language Models | | 85.0 | LFM (w/ Qwen2.5-VL) | 2026-06-03 |
| PerceptionDLM: Parallel Region Perception with Multimodal Diffusion Language Models | ✓ Link | 85.0 | PerceptionDLM-Base (8B) | 2026-06-17 |
| Unlocking Dense Metric Depth Estimation in VLMs | ✓ Link | 84.6 | DepthVLM-8B | 2026-05-15 |
| Qwen3-VL Technical Report | ✓ Link | 83.9 | Qwen3-VL-4B-Instruct | 2025-11-26 |
| Qwen2.5-VL Technical Report | ✓ Link | 83.5 | Qwen2.5-VL-7B | 2025-02-19 |
| Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution | ✓ Link | 83.0 | Qwen2-VL-7B | 2024-09-18 |
| LLaDA2.0-Uni: Unifying Multimodal Understanding and Generation with Diffusion Large Language Model | ✓ Link | 81.5 | LLaDA2.0-Uni (16B) | 2026-04-22 |
| LLaDA-UI: Bringing Block-wise Diffusion to Vision-Language GUI Agents | ✓ Link | 80.5 | LLaDA-UI (pre-training) | 2026-09-09 |
| LLaVA-NeXT: Improved reasoning, OCR, and world knowledge | ✓ Link | 79.3 | LLaVA-NeXT-34B | 2024-01-30 |
| Janus-Pro: Unified Multimodal Understanding and Generation with Data and Model Scaling | ✓ Link | 79.2 | Janus-Pro-7B | 2025-01-29 |
| InfiniteVL: Synergizing Linear and Sparse Attention for Highly-Efficient, Unlimited-Input Vision-Language Models | ✓ Link | 79.0 | InfiniteVL-4B (extra data) | 2025-12-09 |
| Ovis: Structural Embedding Alignment for Multimodal Large Language Model | ✓ Link | 78.4 | Ovis-Qwen1.5-14B | 2024-05-31 |
| MiniCPM-V: A GPT-4V Level MLLM on Your Phone | ✓ Link | 77.2 | MiniCPM-Llama3-V 2.5 | 2024-08-03 |
| What matters when building vision-language models? | ✓ Link | 76.8 | Idefics2 (w/o image split) | 2024-05-03 |
| Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution | ✓ Link | 74.9 | Qwen2-VL-2B | 2024-09-18 |
| NeuPAT: Neuron-aware Plasticity Allocation Tuning for Language-Preserving MLLMs | | 74.31 | NeuPAT (Qwen2.5-14B-Instruct) (extra data) | 2026-08-08 |
| LION: Empowering Multimodal Large Language Model with Dual-Level Visual Knowledge | ✓ Link | 73.4 | LION-12B | 2023-11-20 |
| Token-level Response-visual Attention Guidance for Multimodal LLMs Knowledge Distillation | ✓ Link | 70.8 | TRAG (Qwen2.5-1.5B) | 2026-07-01 |
| Janus: Decoupling Visual Encoding for Unified Multimodal Understanding and Generation | ✓ Link | 69.4 | Janus | 2024-10-17 |
| On the Design Fundamentals of Pixel Text Representation Learning | ✓ Link | 67.87 | Pixel Linguist II + Qwen2.5-7B-Instruct (90% visual tokens) | 2026-09-01 |
| RubiCap: Rubric-Guided Reinforcement Learning for Dense Image Captioning | | 67.61 | RubiCap-7B (PixMoCap) | 2026-03-10 |
| Improved Baselines with Visual Instruction Tuning | ✓ Link | 64.3 | LLaVA-1.5-7B | 2023-10-05 |
| Video-LLaVA: Learning United Visual Representation by Alignment Before Projection | ✓ Link | 60.9 | Video-LLaVA-7B | 2023-11-16 |