Browse › Computer Vision › Visual Question Answering › MM-Vet
Visual Question Answering archive 2025-07-28
MM-Vet Benchmark (Visual Question Answering)
MLLM Leaderboard
The archive carries no text for this table; the description above is the archive's text for the task Visual Question Answering. archive 2025-07-28
Over time archive 2025-07-28
The chart needs JavaScript; the table below carries every value.
Direction inferred from the metric name, not from the archive: GPT-4 score (higher is better), Params (lower is better). Points are placed at the row's paper date; 223 of 231 rows carry one.
Results archive 2025-07-28
Archive rows end at the archive snapshot, 2025-07-28: no result published after that date is in this table. Rank is the archive's row order at that snapshot; not re-ranked here. Metric values are the archive's strings. Column headers sort the table in your browser; each row keeps its archive rank.
| Paper | Code | Ran Syntology | Report | ||||||
|---|---|---|---|---|---|---|---|---|---|
| 1 | gemini-2.0-flash-exp | 81.2±0.4 | – | – | – | not matched | report | ||
| 2 | gemini-exp-1206 | 78.1±0.2 | – | – | – | not matched | report | ||
| 3 | Gemini 1.5 Pro (gemini-1.5-pro-002) | 76.9±0.1 | – | Paper | Code | 2024 | linked, not harvested | report | |
| 4 | MMCTAgent (GPT-4 + GPT-4V) | 74.24 | – | Paper | – | 2024 | no code linked | report | |
| 5 | Claude 3.5 Sonnet (claude-3-5-sonnet-20240620) | 74.2±0.2 | – | Paper | – | 2024 | no code linked | report | |
| 6 | Qwen2-VL-72B | 74.0 | – | Paper | Code | 2024 | 8 of 12 ran · 4 unverified | report | |
| 7 | InternVL2.5-78B | 72.3 | 78B | – | Paper | Code | 2024 | 1 of 9 ran · 8 unverified | report |
| 8 | GPT-4o +text rationale +IoT | 72.2 | – | Paper | – | 2024 | no code linked | report | |
| 9 | Lyra-Pro | 71.4 | 74B | – | Paper | Code | 2024 | 3 of 19 ran · 16 unverified | report |
| 10 | GLM-4V-Plus | 71.1 | – | Paper | Code | 2024 | 12 of 17 ran · 5 unverified | report | |
| 11 | Phantom-7B | 70.8 | – | Paper | Code | 2024 | 1 of 4 ran · 3 unverified | report | |
| 12 | GPT-4o (gpt-4o-2024-05-13) | 69.3±0.1 | – | Paper | Code | 2023 | 2 of 5 ran · 3 unverified | report | |
| 13 | InternVL2.5-38B | 68.8 | 38B | – | Paper | Code | 2024 | 1 of 9 ran · 8 unverified | report |
| 14 | gpt-4o-mini-2024-07-18 | 68.6±0.1 | – | Paper | Code | 2023 | 2 of 5 ran · 3 unverified | report | |
| 15 | GPT-4V | 67.7±0.3 | – | Paper | Code | 2023 | 2 of 5 ran · 3 unverified | report | |
| 16 | GPT-4V-Turbo-detail:high | 67.6±0.1 | – | Paper | Code | 2023 | 2 of 5 ran · 3 unverified | report | |
| 17 | Qwen-VL-Max | 66.6±0.5 | – | Paper | Code | 2023 | 0 of 2 ran · 2 unverified | report | |
| 18 | Gemini 1.5 Pro (gemini-1.5-pro) | 65.8±0.1 | – | Paper | Code | 2024 | linked, not harvested | report | |
| 19 | InternVL2-26B (SGP, token ratio 64%) | 65.60 | 26B | – | Paper | Code | 2024 | 4 of 5 ran · 1 unverified | report |
| 20 | Baichuan-Omni (7B) | 65.4 | – | Paper | Code | 2024 | linked, not harvested | report | |
| 21 | InternVL2.5-26B | 65.0 | 26B | – | Paper | Code | 2024 | 1 of 9 ran · 8 unverified | report |
| 22 | Qwen2-VL-7B (finetuned on GAP-VQA train) | 64.954 | – | Paper | – | 2024 | no code linked | report | |
| 23 | InternVL2-Llama3-76B | 64.4 | – | – | – | not matched | report | ||
| 24 | Gemini 1.0 Pro Vision (gemini-pro-vision) | 64.3±0.4 | – | Paper | Code | 2023 | linked, not harvested | report | |
| 25 | GLM4 Vision | 63.9 | – | Paper | Code | 2023 | linked, not harvested | report | |
| 26 | LLaVA-OneVision-72B | 63.7 | – | Paper | Code | 2024 | linked, not harvested | report | |
| 27 | Lyra-Base | 63.5 | 9B | – | Paper | Code | 2024 | 3 of 19 ran · 16 unverified | report |
| 28 | InternVL2-26B (SGP, token ratio 35%) | 63.20 | 26B | – | Paper | Code | 2024 | 4 of 5 ran · 1 unverified | report |
| 29 | InternVL 1.5 | 62.8 | 26B | – | Paper | Code | 2024 | linked, not harvested | report |
| 30 | InternVL2.5-8B | 62.8 | 8B | – | Paper | Code | 2024 | 1 of 9 ran · 8 unverified | report |
| 31 | MAmmoTH-VL-8B | 62.3 | – | Paper | Code | 2024 | linked, not harvested | report | |
| 32 | Qwen2-VL-7B | 62.0 | – | Paper | Code | 2024 | 8 of 12 ran · 4 unverified | report | |
| 33 | InternVL2-40B | 61.8 | – | – | – | not matched | report | ||
| 34 | Qwen-VL-Plus | 61.1±0.2 | – | Paper | Code | 2023 | 0 of 2 ran · 2 unverified | report | |
| 35 | Mini-Gemini-HD-BS | 60.8 | – | Paper | Code | 2024 | 7 of 8 ran · 1 unverified | report | |
| 36 | InternVL2.5-2B | 60.8 | 2B | – | Paper | Code | 2024 | 1 of 9 ran · 8 unverified | report |
| 37 | MAmmoTH-VL-8B (SI) | 60.6 | – | Paper | Code | 2024 | linked, not harvested | report | |
| 38 | InternVL2.5-4B | 60.6 | 4B | – | Paper | Code | 2024 | 1 of 9 ran · 8 unverified | report |
| 39 | GPT-4V-Turbo-detail:low | 60.2±0.3 | – | Paper | Code | 2023 | 2 of 5 ran · 3 unverified | report | |
| 40 | Mini-Gemini-HD | 59.3 | – | Paper | Code | 2024 | 7 of 8 ran · 1 unverified | report | |
| 41 | Claude 3 Opus (claude-3-opus-20240229) | 58.1±0.1 | – | – | – | not matched | report | ||
| 42 | GLM-4V-9B | 58.0 | – | Paper | Code | 2024 | 12 of 17 ran · 5 unverified | report | |
| 43 | LLaVA-OneVision-7B | 57.5 | – | Paper | Code | 2024 | linked, not harvested | report | |
| 44 | LLaVA-NeXT-34B | 57.4 | 34B | – | – | – | not matched | report | |
| 45 | Meteor | 57.3 | 7B | – | Paper | Code | 2024 | 3 of 8 ran · 5 unverified | report |
| 46 | CROME (Vicuna-13B) | 55.1 | – | Paper | – | 2024 | no code linked | report | |
| 47 | IXC2-4KHD | 54.9 | – | Paper | Code | 2024 | linked, not harvested | report | |
| 48 | Weitu-VL-1.0 | 54.7 | 15B | – | – | – | not matched | report | |
| 49 | TroL-7B | 54.7 | 7B | – | Paper | Code | 2024 | 22 of 30 ran · 8 unverified | report |
| 50 | Mini-Gemini | 53.0 | – | Paper | Code | 2024 | 7 of 8 ran · 1 unverified | report | |
| 51 | CogVLM(Vicuna-7B) | 52.8 | 17B | – | Paper | Code | 2023 | linked, not harvested | report |
| 52 | CogAgent | 52.8 | 18B | – | Paper | Code | 2023 | 12 of 18 ran · 6 unverified | report |
| 53 | Qwen2-VL-2B (finetuned on GAP-VQA train) | 52.43 | – | Paper | – | 2024 | no code linked | report | |
| 54 | InternVL2-26B (SGP, token ratio 9%) | 52.10 | 26B | – | Paper | Code | 2024 | 4 of 5 ran · 1 unverified | report |
| 55 | MM1.5-30B | 52.0 | – | Paper | – | 2024 | no code linked | report | |
| 56 | MiniCPM-Llama3-V-2.5-8B (finetuned on GAP-VQA train) | 51.789 | – | Paper | – | 2024 | no code linked | report | |
| 57 | IXC-2.5-7B | 51.7 | – | Paper | Code | 2024 | 2 of 2 ran · 0 unverified | report | |
| 58 | InternLM-XComposer2 | 51.2 | – | Paper | Code | 2024 | linked, not harvested | report | |
| 59 | Lyra-Mini | 51.2 | 3B | – | Paper | Code | 2024 | 3 of 19 ran · 16 unverified | report |
| 60 | CuMo-7B | 51.0 | 7B | – | Paper | Code | 2024 | 10 of 12 ran · 2 unverified | report |
| 61 | TACO (Qwen2-7B / SigLIP) | 50.9 | – | Paper | Code | 2024 | linked, not harvested | report | |
| 62 | Qwen-VL-Chat (+ SFT (GPT-4V in VLFeedback)) | 50.7 | – | Paper | – | 2024 | 5 of 7 ran · 2 unverified | report | |
| 63 | POINTS-9B | 50.0 | – | Paper | – | 2024 | no code linked | report | |
| 64 | VILA^2-8B | 50.0 | – | Paper | – | 2024 | no code linked | report | |
| 65 | Janus-Pro-7B | 50.0 | – | Paper | Code | 2025 | linked, not harvested | report | |
| 66 | Silkie | 49.9 | 7B | – | Paper | – | 2023 | no code linked | report |
| 67 | Silkie (Qwen-VL-Chat + DPO w/ VLFeedback) | 49.9 | – | Paper | – | 2024 | 5 of 7 ran · 2 unverified | report | |
| 68 | Qwen2-VL-2B | 49.5 | – | Paper | Code | 2024 | 8 of 12 ran · 4 unverified | report | |
| 69 | FlashSloth-HD | 49.0 | 3.2B | – | Paper | Code | 2024 | 7 of 7 ran · 0 unverified | report |
| 70 | InternVL 1.2 | 48.9 | 40B | – | Paper | Code | 2024 | linked, not harvested | report |
| 71 | SEA-PRIME (Vicuna-13B) | 48.8 | – | Paper | – | 2024 | no code linked | report | |
| 72 | InternVL2.5-1B | 48.8 | 1B | – | Paper | Code | 2024 | 1 of 9 ran · 8 unverified | report |
| 73 | MM1-30B-Chat | 48.7 | – | Paper | – | 2024 | no code linked | report | |
| 74 | SETOKIM (13B) | 48.7 | 13B | – | Paper | – | 2024 | no code linked | report |
| 75 | Emu2-Chat | 48.5 | 37B | – | Paper | Code | 2023 | 3 of 4 ran · 1 unverified | report |
| 76 | MG-LLaVA(34B) | 48.5 | – | Paper | Code | 2024 | 1 of 1 ran · 0 unverified | report | |
| 77 | SPHINX-Plus | 47.9 | – | Paper | Code | 2024 | 3 of 3 ran · 0 unverified | report | |
| 78 | ConvLLaVA | 45.9 | 7B | – | Paper | Code | 2024 | 7 of 10 ran · 3 unverified | report |
| 79 | VILA-13B | 45.7 | – | Paper | Code | 2023 | linked, not harvested | report | |
| 80 | TACO (LLaMA3-8B / SigLIP) | 45.7 | – | Paper | Code | 2024 | linked, not harvested | report | |
| 81 | HPT 1.5 Edge | 45.3 | – | – | – | not matched | report | ||
| 82 | TACO (LLaMA3-8B / CLIP) | 45.2 | – | Paper | Code | 2024 | linked, not harvested | report | |
| 83 | LLaVA-v1.6 (7B, w/ STIC) | 45.0 | 7B | – | Paper | Code | 2024 | 3 of 4 ran · 1 unverified | report |
| 84 | H2OVL-Mississippi-2B | 44.7 | – | Paper | – | 2024 | no code linked | report | |
| 85 | PIIP-LLaVA (Vicuna-7B, ConvNeXt-L, CLIP-L ) | 44.7 | 7B | – | Paper | Code | 2025 | linked, not harvested | report |
| 86 | MM-ReAct-GPT-4 | 44.6±0.2 | – | Paper | Code | 2023 | 0 of 4 ran · 4 unverified | report | |
| 87 | Imp-4B | 44.6 | 4B | – | Paper | Code | 2024 | 6 of 7 ran · 1 unverified | report |
| 88 | LLaVA-Next-Mistral-7b (+ DPO w/ VLFeedback) | 44.2 | – | Paper | – | 2024 | 5 of 7 ran · 2 unverified | report | |
| 89 | MGM-7B+RP | 44.1 | – | Paper | – | 2024 | no code linked | report | |
| 90 | LLaVA-Next-Vicuna-7b (+ DPO w/ VLFeedback) | 44.1 | – | Paper | – | 2024 | 5 of 7 ran · 2 unverified | report | |
| 91 | VW-LMM | 44.0 | – | Paper | Code | 2024 | linked, not harvested | report | |
| 92 | MoAI | 43.7 | 7B | – | Paper | Code | 2024 | 6 of 7 ran · 1 unverified | report |
| 93 | MM1-3B-Chat | 43.7 | – | Paper | – | 2024 | no code linked | report | |
| 94 | MM1.5-3B-MoE | 43.7 | – | Paper | – | 2024 | no code linked | report | |
| 95 | Imp-3B | 43.3 | 3B | – | Paper | Code | 2024 | 6 of 7 ran · 1 unverified | report |
| 96 | ShareGPT4V-13B | 43.1 | 13B | – | Paper | Code | 2023 | linked, not harvested | report |
| 97 | Mini-Gemini (+MoCa) | 42.9 | – | Paper | Code | 2024 | 12 of 12 ran · 0 unverified | report | |
| 98 | MM1.5-7B | 42.2 | – | Paper | – | 2024 | no code linked | report | |
| 99 | MM1-7B-Chat | 42.1 | – | Paper | – | 2024 | no code linked | report | |
| 100 | FlashSloth | 41.9 | 3.2B | – | Paper | Code | 2024 | 7 of 7 ran · 0 unverified | report |
| 101 | DeepSeek-VL | 41.5 | – | Paper | Code | 2024 | 7 of 11 ran · 4 unverified | report | |
| 102 | LaVA1.5-13B-BPO | 41.4 | 13B | – | Paper | – | 2024 | no code linked | report |
| 103 | ASMv2 | 41.3 | – | Paper | Code | 2024 | 7 of 8 ran · 1 unverified | report | |
| 104 | FocusLLaVA | 41.3 | – | Paper | – | 2024 | no code linked | report | |
| 105 | SeVa-13B | 41.0 | – | Paper | Code | 2024 | 9 of 11 ran · 2 unverified | report | |
| 106 | MM1.5-3B | 41.0 | – | Paper | – | 2024 | no code linked | report | |
| 107 | LLaVA-1.5-7B (VG-S) | 40.4 | – | Paper | Code | 2024 | 0 of 15 ran · 15 unverified | report | |
| 108 | CoLLaVO | 40.3 | 7B | – | Paper | Code | 2024 | 5 of 6 ran · 1 unverified | report |
| 109 | SPHINX-2k | 40.2 | – | Paper | Code | 2023 | 3 of 3 ran · 0 unverified | report | |
| 110 | LLaVA-1.5 (LVIS-Instrcut4V) | 40.2 | 13B | – | Paper | Code | 2023 | linked, not harvested | report |
| 111 | mPLUG-Owl3 | 40.1 | – | Paper | Code | 2024 | linked, not harvested | report | |
| 112 | Mono-InternVL-2B | 40.1 | 2B | – | Paper | – | 2024 | no code linked | report |
| 113 | LLaVA1.5-13B-MDA | 39.90 | 13B | – | Paper | – | 2024 | no code linked | report |
| 114 | LLaVA-VT (Vicuna-13B) | 39.8 | – | Paper | Code | 2024 | 8 of 8 ran · 0 unverified | report | |
| 115 | MM1.5-1B-MoE | 39.8 | – | Paper | – | 2024 | no code linked | report | |
| 116 | Janus-Pro-1B | 39.8 | – | Paper | Code | 2025 | linked, not harvested | report | |
| 117 | SQ-LLaVA∗ | 39.7 | – | Paper | Code | 2024 | 4 of 6 ran · 2 unverified | report | |
| 118 | OmniFusion (grid split + ruDocVQA) | 39.40 | – | Paper | – | 2024 | no code linked | report | |
| 119 | DeepStack-L-HD (Vicuna-13B) | 39.3 | – | Paper | – | 2024 | no code linked | report | |
| 120 | LAF-13B | 38.9 | – | Paper | – | 2024 | no code linked | report | |
| 121 | InfiMM-HD | 38.9 | – | Paper | – | 2024 | no code linked | report | |
| 122 | InternLM-XC2 + MMDU-45k | 38.8 | – | Paper | Code | 2024 | 7 of 13 ran · 6 unverified | report | |
| 123 | LLaVA-1.5-7B (DC-S) | 38.5 | – | Paper | Code | 2024 | 0 of 15 ran · 15 unverified | report | |
| 124 | LayoutLMv3+ConvNeXt+CLIP | 38.4 | 7.9B | – | Paper | Code | 2024 | linked, not harvested | report |
| 125 | VOLCANO 13B | 38.0 | 13B | – | Paper | Code | 2023 | 9 of 10 ran · 1 unverified | report |
| 126 | LLaVA-1.5+MMInstruct (Vicuna-13B) | 37.9 | – | Paper | Code | 2024 | 2 of 2 ran · 0 unverified | report | |
| 127 | LLaVA-1.5-13B (+CSR) | 37.8 | 13B | – | Paper | Code | 2024 | 1 of 1 ran · 0 unverified | report |
| 128 | LLaVA-1.5-LLaMA3-8B | 37.8 | – | Paper | – | 2024 | no code linked | report | |
| 129 | LLaVA-1.5 + DenseFusion-1M (Vicuna-7B) | 37.8 | – | Paper | Code | 2024 | 2 of 4 ran · 2 unverified | report | |
| 130 | ShareGPT4V-7B | 37.6 | 7B | – | Paper | Code | 2023 | linked, not harvested | report |
| 131 | LLaVA-1.5+CoS | 37.6 | – | Paper | Code | 2024 | 3 of 3 ran · 0 unverified | report | |
| 132 | LLaVA-COCO-13B | 37.5 | – | Paper | – | 2024 | no code linked | report | |
| 133 | LLaVA-S^2 + DenseFusion-1M (Vicuna-7B) | 37.5 | – | Paper | Code | 2024 | 2 of 4 ran · 2 unverified | report | |
| 134 | MM1.5-1B | 37.4 | – | Paper | – | 2024 | no code linked | report | |
| 135 | Dynamic-LLaVA-13B | 37.3 | – | Paper | Code | 2024 | 13 of 18 ran · 5 unverified | report | |
| 136 | SeVa-7B | 37.2 | – | Paper | Code | 2024 | 9 of 11 ran · 2 unverified | report | |
| 137 | SoM-LLaVA-1.5-T | 37.2 | – | Paper | Code | 2024 | linked, not harvested | report | |
| 138 | Emu3 | 37.2 | – | Paper | Code | 2024 | 3 of 6 ran · 3 unverified | report | |
| 139 | LLaVA-Instruct (Vicuna-1.5-13B) | 37.1 | – | Paper | Code | 2024 | linked, not harvested | report | |
| 140 | ILLUME | 37.0 | – | Paper | – | 2024 | no code linked | report | |
| 141 | LLaVA1.5-7B-BPO | 36.8 | 7B | – | Paper | – | 2024 | no code linked | report |
| 142 | LLaVA-1.5-13B (+ MMFuser) | 36.6 | – | Paper | Code | 2024 | linked, not harvested | report | |
| 143 | CaMML-13B | 36.4 | – | Paper | Code | 2024 | linked, not harvested | report | |
| 144 | LLaVA-65B (Data Mixing) | 36.4 | – | Paper | Code | 2023 | 2 of 2 ran · 0 unverified | report | |
| 145 | LLaVA-1.5-13B | 36.3±0.2 | 13B | – | Paper | Code | 2023 | 6 of 9 ran · 3 unverified | report |
| 146 | Emu-14B | 36.3±0.3 | 14B | – | Paper | Code | 2023 | 0 of 2 ran · 2 unverified | report |
| 147 | mPLUG-Owl2 | 36.3±0.1 | 7B | – | Paper | Code | 2023 | 1 of 3 ran · 2 unverified | report |
| 148 | Vary-base | 36.2 | – | Paper | Code | 2023 | linked, not harvested | report | |
| 149 | StableLLaVA | 36.1 | – | Paper | Code | 2023 | 6 of 7 ran · 1 unverified | report | |
| 150 | DreamLLM-7B | 35.9 | 7B | – | Paper | Code | 2023 | 3 of 3 ran · 0 unverified | report |
| 151 | MoE-LLaVA-2.7B×4-Top2 | 35.9 | – | Paper | Code | 2024 | 0 of 2 ran · 2 unverified | report | |
| 152 | SoM-LLaVA-1.5 | 35.9 | – | Paper | Code | 2024 | linked, not harvested | report | |
| 153 | Dragonfly (Llama3-8B) | 35.9 | – | Paper | Code | 2024 | 1 of 1 ran · 0 unverified | report | |
| 154 | Ferret-v2-13B | 35.7 | – | Paper | Code | 2024 | 4 of 5 ran · 1 unverified | report | |
| 155 | AlignGPT (Vicuna-13B) | 35.6 | – | Paper | – | 2024 | no code linked | report | |
| 156 | LLaVA-HR-X | 35.5 | – | Paper | Code | 2024 | 5 of 7 ran · 2 unverified | report | |
| 157 | SQ-LLaVA | 35.5 | – | Paper | Code | 2024 | 4 of 6 ran · 2 unverified | report | |
| 158 | LOVA³ | 35.2 | 7B | – | Paper | Code | 2024 | 7 of 10 ran · 3 unverified | report |
| 159 | LLaVA-InternLM2-7B-ViT + MoSLoRA | 35.2 | – | Paper | Code | 2024 | 4 of 6 ran · 2 unverified | report | |
| 160 | InternLM2+ViT (QMoSLoRA) | 35.2 | – | Paper | Code | 2024 | 4 of 6 ran · 2 unverified | report | |
| 161 | LLaVA1.5-7B-MDA | 35.20 | 7B | – | Paper | – | 2024 | no code linked | report |
| 162 | Mipha-3B+ | 35.1 | – | Paper | – | 2024 | no code linked | report | |
| 163 | LLaVA-Plus-13B (All Tools, V1.3, 336px) | 35.0±0.0 | 13B | – | Paper | Code | 2023 | 2 of 2 ran · 0 unverified | report |
| 164 | Merlin | 34.9 | – | Paper | – | 2023 | no code linked | report | |
| 165 | Arcana | 34.8 | – | Paper | – | 2024 | no code linked | report | |
| 166 | INF-LLaVA | 34.5 | – | Paper | Code | 2024 | 4 of 4 ran · 0 unverified | report | |
| 167 | LLaVA-1.5+MMInstruct (Vicuna-7B) | 34.4 | – | Paper | Code | 2024 | 2 of 2 ran · 0 unverified | report | |
| 168 | Janus | 34.3 | – | Paper | Code | 2024 | linked, not harvested | report | |
| 169 | LLaVA-TokenPacker (Vicuna-13B) | 34.1 | – | Paper | Code | 2024 | 1 of 1 ran · 0 unverified | report | |
| 170 | γ-MoD-LLaVA-HR | 34.0 | – | Paper | – | 2024 | no code linked | report | |
| 171 | LLaVA-1.5-7B (CSR) | 33.9 | 7B | – | Paper | Code | 2024 | 1 of 1 ran · 0 unverified | report |
| 172 | DynMOE-LLaVA | 33.6 | – | Paper | Code | 2024 | 2 of 2 ran · 0 unverified | report | |
| 173 | Imp-2B | 33.5 | 2B | – | Paper | Code | 2024 | 6 of 7 ran · 1 unverified | report |
| 174 | InfMLLM-7B-Chat | 33.4 | – | Paper | Code | 2023 | linked, not harvested | report | |
| 175 | Video-LaVIT | 33.2 | 7B | – | Paper | Code | 2024 | 3 of 5 ran · 2 unverified | report |
| 176 | LLaVA-Instruct (Vicuna-1.5-7B) | 32.9 | – | Paper | Code | 2024 | linked, not harvested | report | |
| 177 | VisionZip (Retain 128 Tokens, fine-tuning) | 32.9 | – | Paper | Code | 2024 | 1 of 11 ran · 10 unverified | report | |
| 178 | Uni-MoE | 32.8 | – | Paper | Code | 2024 | 6 of 8 ran · 2 unverified | report | |
| 179 | VL-Mamba (Mamba LLM-2.8B) | 32.6 | – | Paper | – | 2024 | no code linked | report | |
| 180 | LLaVA-v1.5 (7B, w/ STIC) | 32.6 | 7B | – | Paper | Code | 2024 | 3 of 4 ran · 1 unverified | report |
| 181 | VisionZip (Retain 192 Tokens, fine-tuning) | 32.6 | – | Paper | Code | 2024 | 1 of 11 ran · 10 unverified | report | |
| 182 | VisionZip (Retain 128 Tokens) | 32.6 | – | Paper | Code | 2024 | 1 of 11 ran · 10 unverified | report | |
| 183 | LLaVA-v1.5 (+MoCa) | 32.2 | – | Paper | Code | 2024 | 12 of 12 ran · 0 unverified | report | |
| 184 | Dynamic-LLaVA-7B | 32.2 | – | Paper | Code | 2024 | 13 of 18 ran · 5 unverified | report | |
| 185 | Mipha-3B | 32.1 | 3B | – | Paper | Code | 2024 | linked, not harvested | report |
| 186 | VOLCANO 7B | 32.0 | 7B | – | Paper | Code | 2023 | 9 of 10 ran · 1 unverified | report |
| 187 | Video-LLaVA | 32.0 | – | Paper | Code | 2023 | 4 of 7 ran · 3 unverified | report | |
| 188 | TinyLLaVA-share-Sig-Ph | 32.0 | 3.1B | – | Paper | Code | 2024 | 5 of 10 ran · 5 unverified | report |
| 189 | LLaVA-VT (Vicuna-7B) | 31.8 | – | Paper | Code | 2024 | 8 of 8 ran · 0 unverified | report | |
| 190 | LRV-Instruction-7B | 31.7±0.1 | 7B | – | Paper | Code | 2023 | linked, not harvested | report |
| 191 | VisionZip (Retain 192 Tokens) | 31.7 | – | Paper | Code | 2024 | 1 of 11 ran · 10 unverified | report | |
| 192 | VisionZip (Retain 64 Tokens) | 31.7 | – | Paper | Code | 2024 | 1 of 11 ran · 10 unverified | report | |
| 193 | LLaVA-1.5-7B (+ SIMA) | 31.6 | 7B | – | Paper | Code | 2024 | 5 of 5 ran · 0 unverified | report |
| 194 | LLaMA-Adapter v2-7B | 31.4±0.1 | 7B | – | Paper | Code | 2023 | 0 of 1 ran · 1 unverified | report |
| 195 | MiCo-Chat-7B | 31.4 | – | Paper | Code | 2024 | linked, not harvested | report | |
| 196 | LLaVA-1.5-7B + TeamLoRA | 31.2 | – | Paper | Code | 2024 | linked, not harvested | report | |
| 197 | LLaVA-1.5-7B | 31.1±0.2 | 7B | – | Paper | Code | 2023 | 6 of 9 ran · 3 unverified | report |
| 198 | RoboCodeX-13B | 31.0 | – | Paper | – | 2024 | no code linked | report | |
| 199 | HyperLLaVA | 31.0 | – | Paper | Code | 2024 | linked, not harvested | report | |
| 200 | FAST (Vicuna-7B) | 31.0 | – | Paper | Code | 2024 | 3 of 3 ran · 0 unverified | report | |
| 201 | JanusFlow | 30.9 | – | Paper | Code | 2024 | linked, not harvested | report | |
| 202 | AlignGPT (Vicuna-7B) | 30.8 | – | Paper | – | 2024 | no code linked | report | |
| 203 | LLaVolta | 30.7 | – | Paper | Code | 2024 | 7 of 9 ran · 2 unverified | report | |
| 204 | LLaVA-AlignedVQ | 30.7 | – | Paper | – | 2024 | no code linked | report | |
| 205 | LLaVA-1.5-HACL | 30.4 | 7.3B | – | Paper | Code | 2023 | 3 of 6 ran · 3 unverified | report |
| 206 | MaVEn | 30.4 | 7.2B | – | Paper | – | 2024 | no code linked | report |
| 207 | VisionZip (Retain 64 Tokens, fine-tuning) | 30.2 | – | Paper | Code | 2024 | 1 of 11 ran · 10 unverified | report | |
| 208 | H2OVL-Mississippi-0.8B | 30.0 | – | Paper | – | 2024 | no code linked | report | |
| 209 | RoboMamba | 29.7 | – | Paper | – | 2024 | no code linked | report | |
| 210 | LLaVA-TokenPacker (Vicuna-7B) | 29.6 | – | Paper | Code | 2024 | 1 of 1 ran · 0 unverified | report | |
| 211 | OneLLM-7B | 29.1 | 7B | – | Paper | Code | 2023 | 4 of 4 ran · 0 unverified | report |
| 212 | LLaVA-OneVision-0.5B | 29.1 | – | Paper | Code | 2024 | linked, not harvested | report | |
| 213 | Vary-toy | 29.0 | 1.8B | – | Paper | – | 2024 | no code linked | report |
| 214 | LLaVA-Phi | 28.9 | – | Paper | Code | 2024 | 5 of 9 ran · 4 unverified | report | |
| 215 | MM-ReAct-GPT-3.5 | 27.9±0.1 | – | Paper | Code | 2023 | 0 of 4 ran · 4 unverified | report | |
| 216 | MMAR-7B | 27.80 | – | Paper | – | 2024 | no code linked | report | |
| 217 | SEAL (7B) | 27.7 | 7B | – | Paper | Code | 2023 | 5 of 5 ran · 0 unverified | report |
| 218 | LLaVA-Plus-7B (All Tools) | 27.5±0.3 | 7B | – | Paper | Code | 2023 | 2 of 2 ran · 0 unverified | report |
| 219 | OtterHD-8B | 26.3 | 8B | – | Paper | Code | 2023 | 11 of 14 ran · 3 unverified | report |
| 220 | TGA-7B | 25.6 | – | Paper | – | 2024 | no code linked | report | |
| 221 | OpenFlamingo-9B (MPT-7B) | 24.8±0.2 | 9B | – | Paper | Code | 2023 | linked, not harvested | report |
| 222 | Otter-9B (MPT-7B) | 24.7±0.3 | 9B | – | Paper | Code | 2023 | linked, not harvested | report |
| 223 | Otter-9B (LLaMA) | 24.6±0.2 | 9B | – | Paper | Code | 2023 | linked, not harvested | report |
| 224 | MiniGPT-4-14B | 24.4±0.4 | 14B | – | Paper | Code | 2023 | linked, not harvested | report |
| 225 | LinVT | 23.5 | – | Paper | Code | 2024 | 4 of 12 ran · 8 unverified | report | |
| 226 | BLIP-2-12B | 22.4±0.2 | 12B | – | Paper | Code | 2023 | 4 of 8 ran · 4 unverified | report |
| 227 | MiniGPT-4-8B | 22.1±0.1 | 8B | – | Paper | Code | 2023 | linked, not harvested | report |
| 228 | OpenFlamingo-9B (LLaMA-7B) | 21.8±0.1 | 9B | – | Paper | Code | 2023 | linked, not harvested | report |
| 229 | Xmodel-VLM (Xmodel-LM 1.1B) | 21.8 | – | Paper | Code | 2024 | linked, not harvested | report | |
| 230 | TextBind | 19.4 | – | Paper | Code | 2023 | linked, not harvested | report | |
| 231 | MMAR-0.5B | 18.49 | – | Paper | – | 2024 | no code linked | report |
All 231 rows shown. 223 link to a paper page on this site; 0 are marked as using additional training data in the archive. No GitHub stars are tracked; "Code" is the first repository the archive lists for the row. The archive carries no row tags, review links or community-submitted rows for this table; none are shown. archive 2025-07-28
Syntology Ran reads "N of M ran · U unverified": of the M code samples Syntology harvested from repositories linked to that row's paper (joined by arXiv id), N executed on a synthesized input and the other U = M−N are unverified (harvested, no recorded run). It counts code from repositories linked to that row's paper, not this result: the row's number was not reproduced and nothing here is a correctness claim. The other cell texts mean no graph line for the row: "linked, not harvested" (the archive links code, Syntology has not harvested it), "no code linked" (no code link in the archive), "not matched" (the row's paper URL matched no paper on this site). 120 rows have a graph line, from 73 distinct papers; 111 rows (67 papers) have at least one sample that ran. Counting each paper once: Syntology ran 341 of 525 samples; 184 unverified. Separately, 180 of those 525 are pointer-only (licence): the site points at that code rather than redistributing it, a licence property recorded for ran and unverified samples alike; each cell's tooltip carries the row's own pointer-only count. Read from the graph 2026-09-24. Per-sample status is on the paper page.
Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections