Retrieval heads stay sparse and shared across MLLM modalities
Ranking training on those heads raises Qwen3-VL-8B page Recall@1 on MMDocIR by 6.03 points.
44
AI model
Qwen3-VL-8B
- Mentions · 7d0
- Mentions · 30d4
- All mentions13
- First seenSep 15, 2026
- Last mentioned20d ago
Timeline
13 articles · newest firstOpenVLHarness adds tools and memory for visual agent reasoning
On 22 benchmarks, average scores rise 9.7–12.6 points over matched no-tool baselines across eight backbones.
42
OpenVLHarness adds tool use and memory for visual agent reasoning
On 22 benchmarks it lifts matched no-tool baselines by 9.7 to 12.6 points across eight backbones.
48
Twice Upon a Time: Complementary Representations for Time-Series Reasoning
A Qwen3-VL-8B dual-view time-series model lifts TSRBench accuracy from 44.5% to 53.2%.
46
FastBench: Can Streaming VLMs Perceive High-Dynamic Real-World Streams?
FastBench shows streaming VLMs miss fast events; Gemini-3.5-Flash scores only 50.7%.
44
OneSearch-VL: Unified Multimodal Deep Research Agent for Image and Video
OneSearch-VL-8B beats tool-using Qwen3-VL-8B by up to 20.2 points on new visual research benchmarks.
58
Foresight: planning future perception in streaming VLMs without retraining
FORESIGHT lets frozen Qwen3-VL-8B plan future perception in streaming video without retraining.
52
Ranking-Aware Prompt Optimization for Multimodal Clinical Diagnosis
Ranking-aware prompt evolution lifts clinical AUROC on Qwen3-VL-8B and MedGemma-4B versus accuracy-based search.
52
From Given to Gathered Evidence: Agentic Learning for Longitudinal Medical Reasoning
CASE trains compact medical agents to gather longitudinal evidence, with an 8B model outperforming larger frontier systems.
55
MAD-Guard: Controlled Study of Autoregressive Generation versus Direct Decision Interfaces for Closed Multimodal Forensic Tasks
MAD-Guard shows direct decision heads outperform autoregressive decoding for closed multimodal forgery forensics.
42
unsloth/Qwen-Image-2.1-GGUF — new model trending #30 on Hugging Face
Unsloth published GGUF quantized versions of Qwen-Image-2.1, Alibaba's 7B text-to-image and editing model, using Dynamic 2.0 per-layer precision.
38
pottokao/Qwen-Image-2.1-Text-Encoder-Heretic-GGUF — new model trending #30 on Hugging Face
Community GGUF, FP8, and BF16 builds of the Qwen-Image-2.1 text encoder are trending on Hugging Face.
32
Comfy-Org/Qwen-Image-2.1 — new model trending #30 on Hugging Face
Comfy-Org published repackaged Qwen-Image-2.1 weights for ComfyUI, including text-to-image and image-editing workflow templates.
28
Appears with
- Qwenvendor · 2
- GPT-6 Lunaai model · 2
- Qwen-Image-2.1ai model · 2
- Alibabavendor · 2
- GPT-6 Astraai model · 2
- ComfyUIproduct · 2
- Kimi K3ai model · 2
- Qwen-Image-2.1product · 1
- SAM3ai model · 1
- CoTracker3ai model · 1
- Gemini-3.5-Flashai model · 1
- Qwen3-VL-8Bproduct · 1
- Huaweivendor · 1
- Alibabaorganization · 1
- Unslothorganization · 1
- Claude Opus 4.8ai model · 1
Entities are extracted by the model from each article. First seen , last mentioned . Watching keeps it in this browser only (no account); the watchlist page and dashboard alerts use it.