List your company

Real-world QA on RealworldQA

81.8AccuracyM2Note
49.76858.08466.474.716Aug 6, 2024Nov 29, 2024Mar 25, 2025Jul 19, 2025Nov 11, 2025Mar 7, 2026Jul 1, 2026

Evaluation results

MethodLinks
M2NoteBackbone=GPT-5.4, Evolving Mode=Self-Evolving, Mem=28, Len=4622026.0781.8
VanillaBackbone=GPT-5.4, Evolving Mode=Self-Evolving2026.0780.8
M2NoteBackbone=Qwen3-VL-Plus, Evolving Mode=Self-Evolving, Mem=28, Len=4622026.0779.7
CoT+M2NoteBackbone=Qwen3-VL-8B-Instruct, Evolving Mode=Self-Evolving, Mem=32, Len=3152026.0778.8
VanillaBackbone=Qwen3-VL-Plus, Evolving Mode=Self-Evolving2026.0778.2
CoTBackbone=Qwen3-VL-8B-Instruct, Evolving Mode=Self-Evolving2026.0778
M2NoteBackbone (Tuning)=Qwen3-VL-8B-Instruct, Backbone (Tuner)=Qwen3-VL-Plus, Evolving Mode=Cross-Model Evolving, Mem=43, Len=4452026.0777.1
Qwen2.5-VL-72B2026.0475.7
GPT-4o2026.0475.4
M2NoteBackbone=Qwen3-VL-8B-Instruct, Evolving Mode=Self-Evolving, Mem=50, Len=2612026.0775.4
Qwen2.5-VL-7B AIFmethod=Adaptive Information Flow2026.0474.5
VanillaBackbone=Qwen3-VL-8B-Instruct, Evolving Mode=Self-Evolving2026.0773.2
VanillaBackbone (Tuning)=Qwen3-VL-8B-Instruct, Backbone (Tuner)=Qwen3-VL-Plus, Evolving Mode=Cross-Model Evolving2026.0773.2
SwimBird2026.0273.1
S^3-FTBackbone=Qwen3-VL-8B, Mode=Single-view2026.0572.2
DPEBackbone=Qwen3-VL-8B-Instruct, Evolving Mode=Self-Evolving2026.0772.1
LLaVA-OneVision-72BModel Scale=72B2024.0871.9
SAIL-VLModel Size=8B2025.0171.9
Qwen3-VL-8B-Instructreproduced by authors=true2026.0271.8
LFM2-VL-3B# Total Params=3.0B, Evaluation Toolkit=VLMEvalKit2025.1171.37
Qwen3-VL-8BMode=base2026.0571
InternVL2.5-MPOModel Size=8B2025.0170.2
DeepSeekVL-2Model Size=8B2025.0170.2
LLaVA-OneVision2026.0269.9
Qwen2-VLModel Size=8B2025.0169.7
SkiLa2026.0269.3
S^3-FTBackbone=Qwen3-VL-8B, Mode=Cross-viewpoint2026.0569
VisPlayEvolving Mode=Self-Evolving2026.0769
S^3-FTBackbone=Qwen2.5-VL-7B, Mode=Single-view2026.0568.8
Qwen2.5-VL-7B2026.0468.5
Vision-ZeroEvolving Mode=Self-Evolving2026.0768.5
Qwen2.5-VL-7BMode=base2026.0568.2
Qwen2.5-VL-7B-Instruct2026.0267.4
InfiniteVL-4BComplexity=O(1)2025.1267.3
S^3-FTBackbone=Qwen2.5-VL-7B, Mode=Cross-viewpoint2026.0566.7
LLaVA-OneVision-7BModel Scale=7B2024.0866.3
Qwen3-VL-2B# Total Params=2.1B, Evaluation Toolkit=VLMEvalKit2025.1165.75
InternVL3.5-8BRatio=100%, Backbone=InternVL3.5-8B2026.0565.62
Qwen2.5VL-3BComplexity=O(n)2025.1265.4
Qwen2.5-VL-3B# Total Params=3.75B, Evaluation Toolkit=VLMEvalKit2025.1165.23
DeepSeekVL-2Model Size=2B2025.0164.8
InternVL2.5-4BComplexity=O(n)2025.1264.3
F3ARatio=60%, Backbone=InternVL3.5-8B2026.0563.98
SAIL-VLModel Size=2B2025.0163.8
VisionZipRatio=60%, Backbone=InternVL3.5-8B2026.0563.39
CDPrunerRatio=60%, Backbone=InternVL3.5-8B2026.0563.34
DivPruneRatio=60%, Backbone=InternVL3.5-8B2026.0563.26
CDPrunerRatio=40%, Backbone=InternVL3.5-8B2026.0562.88
CASA⊕ Qwen2.5-VLRes.=≤ 896²2025.1262.5
FastVRatio=60%, Backbone=InternVL3.5-8B2026.0562.27
F3ARatio=40%, Backbone=InternVL3.5-8B2026.0562.26
Qwen2.5-VL 3B (reproduced)Res.=≤ 896²2025.1262.2
VisionZipRatio=40%, Backbone=InternVL3.5-8B2026.0561.98
DivPruneRatio=40%, Backbone=InternVL3.5-8B2026.0561.68
GPT-4VVersion=V-Preview2024.0861.4
FastVRatio=40%, Backbone=InternVL3.5-8B2026.0561.03
Qwen2-VLModel Size=2B2025.0160.9
InternVL3.5-2B# Total Params=2.3B, Evaluation Toolkit=VLMEvalKit2025.1160.78
InternVL2.5-MPOModel Size=2B2025.0160.7
LLaVA-1.5-7B AIFmethod=Adaptive Information Flow2026.0460.5
Claude-3.5 Sonnet2026.0460.1
F3ARatio=20%, Backbone=InternVL3.5-8B2026.0559.85
DivPruneRatio=20%, Backbone=InternVL3.5-8B2026.0559.39
SmolVLM2-2BComplexity=O(n)2025.1258.4
PaliGemma2-3BComplexity=O(n)2025.1258.3
VisionZipRatio=20%, Backbone=InternVL3.5-8B2026.0558.07
Phi-3.5-Vision-4BComplexity=O(n)2025.1257.9
SmolVLM2-2.2B# Total Params=2.2B, Evaluation Toolkit=VLMEvalKit2025.1157.5
LLaVA-OneVision-0.5BModel Scale=0.5B2024.0855.6
LLaVA-1.5-7B2026.0455.6
CDPrunerRatio=20%, Backbone=InternVL3.5-8B2026.0555.42
TinyLLaVA-3BComplexity=O(n)2025.1255
FastVRatio=20%, Backbone=InternVL3.5-8B2026.0552.63
MaTVLM-3BComplexity=O(n)2025.1252.3
Cobra-3BComplexity=O(1)2025.1251