Crossmodal-3600
Models describe images in different languages or retrieve an image from a description.
The benchmark uses human-written captions in 36 languages.
- Made by
- Ashish Thapliyal and colleagues
- Released
- May 2022
- Kind
- General visual questions
- Level
- School
- Scored
- Checked automatically
- Answers
- Prediction
- Tasks
- All public
- Measures
- Capability
- Language
- Many languages
- Interaction
- Single turn
- Refresh
- One fixed set
- Run by
- Anyone
- In views
- Vision, Image, video and 3D generation, Many languages
- Last model added
- October 2026
- Scores refreshed
- November 20th, 2026
dnotitia/DNA-VL-STEER-2B tops the leaderboard with 0.71 EmbeddingGemma 2 leads with 0.65
Best score per entry from MTEB’s leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. The best model we track is EmbeddingGemma 2, at 0.65. This view shows only models we track, each at its best.
- 0.71 dnotitia/DNA-VL-STEER-2B
- 0.69 Haon-Chen/e5-omni-7B
- 0.67 royokong/e5-v
- 0.65
EmbeddingGemma 2
- 0.64 BidirLM/BidirLM-Omni-2.5B-Embedding
- 0.61 Tevatron/OmniEmbed-v0.1
- 0.61 LCO-Embedding/LCO-Embedding-Omni-7B
- 0.58 google/siglip-base-patch16-256-multilingual
- 0.56 jinaai/jina-embeddings-v5-omni-small
- 0.53 Haon-Chen/e5-omni-3B
- 0.52 LCO-Embedding/LCO-Embedding-Omni-3B
- 0.50 mongodb/voyage-multimodal-3
- 0.44 google/siglip-large-patch16-384
- 0.44 google/siglip-large-patch16-256
- 0.41 BAAI/bge-visualized-m3
- 0.37 jinaai/jina-embeddings-v5-omni-nano
- 0.36 google/siglip-base-patch16-512
- 0.35 google/siglip-base-patch16-384
- 0.35 google/siglip-base-patch16-224
- 0.35 google/siglip-base-patch16-256
- 0.30 google/siglip-so400m-patch14-384
- 0.23 TIGER-Lab/VLM2Vec-Full
- 0.23 TIGER-Lab/VLM2Vec-LoRA
- 0.18 laion/CLIP-ViT-bigG-14-laion2B-39B-b160k
- 0.18 encord-team/ebind-full
- 0.17 QuanSun/EVA02-CLIP-bigE-14-plus
- 0.16 kakaobrain/align-base
- 0.16 laion/CLIP-ViT-g-14-laion2B-s34B-b88K
- 0.15 laion/CLIP-ViT-H-14-laion2B-s32B-b79K
- 0.15 QuanSun/EVA02-CLIP-bigE-14
- 0.13 laion/CLIP-ViT-L-14-DataComp.XL-s13B-b90K
- 0.13 LanguageBind/LanguageBind_Omni
- 0.13 laion/CLIP-ViT-L-14-laion2B-s32B-b82K
- 0.13 QuanSun/EVA02-CLIP-L-14
- 0.12 laion/CLIP-ViT-B-16-DataComp.XL-s13B-b90K
- 0.12 Salesforce/blip-itm-large-coco
- 0.12 laion/CLIP-ViT-B-32-DataComp.XL-s13B-b90K
- 0.11 laion/CLIP-ViT-B-32-laion2B-s34B-b79K
- 0.11 jinaai/jina-clip-v1
- 0.11 Salesforce/blip-itm-large-flickr
- 0.11 QuanSun/EVA02-CLIP-B-16
- 0.10 Salesforce/blip-itm-base-coco
- 0.09 openai/clip-vit-large-patch14
- 0.08 Salesforce/blip2-opt-2.7b
- 0.08 openai/clip-vit-base-patch16
- 0.08 Salesforce/blip-itm-base-flickr
- 0.08 Salesforce/blip2-opt-6.7b-coco
- 0.08 openai/clip-vit-base-patch32
- 0.08 nomic-ai/nomic-embed-vision-v1.5
- 0.06 BAAI/bge-visualized-base
- 0.00 Salesforce/blip-image-captioning-large
- 0.00 Salesforce/blip-image-captioning-base
- 1 dnotitia/DNA-VL-STEER-2B Text-to-image retrieval 0.71
- 2 Haon-Chen/e5-omni-7B Text-to-image retrieval 0.69
- 3 royokong/e5-v Text-to-image retrieval 0.67
- 4
EmbeddingGemma 2 Text-to-image retrieval 0.65
- 5 BidirLM/BidirLM-Omni-2.5B-Embedding Text-to-image retrieval 0.64
- 6 Tevatron/OmniEmbed-v0.1 Text-to-image retrieval 0.61
- 7 LCO-Embedding/LCO-Embedding-Omni-7B Text-to-image retrieval 0.61
- 8 google/siglip-base-patch16-256-multilingual Text-to-image retrieval 0.58
- 9 jinaai/jina-embeddings-v5-omni-small Text-to-image retrieval 0.56
- 10 Haon-Chen/e5-omni-3B Text-to-image retrieval 0.53
- 11 LCO-Embedding/LCO-Embedding-Omni-3B Text-to-image retrieval 0.52
- 12 mongodb/voyage-multimodal-3 Text-to-image retrieval 0.50
- 13 google/siglip-large-patch16-384 Text-to-image retrieval 0.44
- 14 google/siglip-large-patch16-256 Text-to-image retrieval 0.44
- 15 BAAI/bge-visualized-m3 Text-to-image retrieval 0.41
- 16 jinaai/jina-embeddings-v5-omni-nano Text-to-image retrieval 0.37
- 17 google/siglip-base-patch16-512 Text-to-image retrieval 0.36
- 18 google/siglip-base-patch16-384 Text-to-image retrieval 0.35
- 19 google/siglip-base-patch16-224 Text-to-image retrieval 0.35
- 20 google/siglip-base-patch16-256 Text-to-image retrieval 0.35
- 21 google/siglip-so400m-patch14-384 Text-to-image retrieval 0.30
- 22 TIGER-Lab/VLM2Vec-Full Text-to-image retrieval 0.23
- 23 TIGER-Lab/VLM2Vec-LoRA Text-to-image retrieval 0.23
- 24 laion/CLIP-ViT-bigG-14-laion2B-39B-b160k Text-to-image retrieval 0.18
- 25 encord-team/ebind-full Text-to-image retrieval 0.18
- 26 QuanSun/EVA02-CLIP-bigE-14-plus Text-to-image retrieval 0.17
- 27 kakaobrain/align-base Text-to-image retrieval 0.16
- 28 laion/CLIP-ViT-g-14-laion2B-s34B-b88K Text-to-image retrieval 0.16
- 29 laion/CLIP-ViT-H-14-laion2B-s32B-b79K Text-to-image retrieval 0.15
- 30 QuanSun/EVA02-CLIP-bigE-14 Text-to-image retrieval 0.15
- 31 laion/CLIP-ViT-L-14-DataComp.XL-s13B-b90K Text-to-image retrieval 0.13
- 32 LanguageBind/LanguageBind_Omni Text-to-image retrieval 0.13
- 33 laion/CLIP-ViT-L-14-laion2B-s32B-b82K Text-to-image retrieval 0.13
- 34 QuanSun/EVA02-CLIP-L-14 Text-to-image retrieval 0.13
- 35 laion/CLIP-ViT-B-16-DataComp.XL-s13B-b90K Text-to-image retrieval 0.12
- 36 Salesforce/blip-itm-large-coco Text-to-image retrieval 0.12
- 37 laion/CLIP-ViT-B-32-DataComp.XL-s13B-b90K Text-to-image retrieval 0.12
- 38 laion/CLIP-ViT-B-32-laion2B-s34B-b79K Text-to-image retrieval 0.11
- 39 jinaai/jina-clip-v1 Text-to-image retrieval 0.11
- 40 Salesforce/blip-itm-large-flickr Text-to-image retrieval 0.11
- 41 QuanSun/EVA02-CLIP-B-16 Text-to-image retrieval 0.11
- 42 Salesforce/blip-itm-base-coco Text-to-image retrieval 0.10
- 43 openai/clip-vit-large-patch14 Text-to-image retrieval 0.09
- 44 Salesforce/blip2-opt-2.7b Text-to-image retrieval 0.08
- 45 openai/clip-vit-base-patch16 Text-to-image retrieval 0.08
- 46 Salesforce/blip-itm-base-flickr Text-to-image retrieval 0.08
- 47 Salesforce/blip2-opt-6.7b-coco Text-to-image retrieval 0.08
- 48 openai/clip-vit-base-patch32 Text-to-image retrieval 0.08
- 49 nomic-ai/nomic-embed-vision-v1.5 Text-to-image retrieval 0.08
- 50 BAAI/bge-visualized-base Text-to-image retrieval 0.06
- 51 Salesforce/blip-image-captioning-large Text-to-image retrieval 0.00
- 52 Salesforce/blip-image-captioning-base Text-to-image retrieval 0.00
Source: MTEB’s leaderboard
Source: MTEB’s leaderboard
| # | Entry | Score | Setting | Date | Link | Source |
|---|---|---|---|---|---|---|
| 1 | dnotitia/DNA-VL-STEER-2B open weights, 2.1B | 0.71 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 2 | Haon-Chen/e5-omni-7B open weights, 8.9B | 0.69 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 3 | royokong/e5-v open weights, 8.4B | 0.67 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 4 | | 0.65 | Text-to-image retrieval | Oct 2026 | huggingface.co | Official leaderboard |
| 5 | BidirLM/BidirLM-Omni-2.5B-Embedding open weights, 2.4B | 0.64 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 6 | Tevatron/OmniEmbed-v0.1 open weights, 8.9B | 0.61 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 7 | LCO-Embedding/LCO-Embedding-Omni-7B open weights, 8.9B | 0.61 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 8 | google/siglip-base-patch16-256-multilingual open weights, 0.4B | 0.58 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 9 | jinaai/jina-embeddings-v5-omni-small open weights, 1.6B | 0.56 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 10 | Haon-Chen/e5-omni-3B open weights, 4.7B | 0.53 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 11 | LCO-Embedding/LCO-Embedding-Omni-3B open weights, 4.7B | 0.52 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 12 | mongodb/voyage-multimodal-3 closed weights | 0.50 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 13 | google/siglip-large-patch16-384 open weights, 0.7B | 0.44 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 14 | google/siglip-large-patch16-256 open weights, 0.7B | 0.44 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 15 | BAAI/bge-visualized-m3 open weights, 0.9B | 0.41 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 16 | jinaai/jina-embeddings-v5-omni-nano open weights, 1B | 0.37 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 17 | google/siglip-base-patch16-512 open weights, 0.2B | 0.36 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 18 | google/siglip-base-patch16-384 open weights, 0.2B | 0.35 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 19 | google/siglip-base-patch16-224 open weights, 0.2B | 0.35 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 20 | google/siglip-base-patch16-256 open weights, 0.2B | 0.35 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 21 | google/siglip-so400m-patch14-384 open weights, 0.9B | 0.30 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 22 | TIGER-Lab/VLM2Vec-Full open weights, 4.1B | 0.23 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 23 | TIGER-Lab/VLM2Vec-LoRA open weights, 4.2B | 0.23 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 24 | laion/CLIP-ViT-bigG-14-laion2B-39B-b160k open weights, 2.5B | 0.18 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 25 | encord-team/ebind-full open weights, 1.8B | 0.18 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 26 | QuanSun/EVA02-CLIP-bigE-14-plus open weights, 5B | 0.17 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 27 | kakaobrain/align-base open weights, 0.2B | 0.16 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 28 | laion/CLIP-ViT-g-14-laion2B-s34B-b88K open weights, 1.4B | 0.16 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 29 | laion/CLIP-ViT-H-14-laion2B-s32B-b79K open weights, 1B | 0.15 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 30 | QuanSun/EVA02-CLIP-bigE-14 open weights, 4.7B | 0.15 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 31 | laion/CLIP-ViT-L-14-DataComp.XL-s13B-b90K open weights, 0.4B | 0.13 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 32 | LanguageBind/LanguageBind_Omni open weights, 1.2B | 0.13 | Text-to-image retrieval | github.com | Official leaderboard | |
| 33 | laion/CLIP-ViT-L-14-laion2B-s32B-b82K open weights, 0.4B | 0.13 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 34 | QuanSun/EVA02-CLIP-L-14 open weights, 0.4B | 0.13 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 35 | laion/CLIP-ViT-B-16-DataComp.XL-s13B-b90K open weights, 0.2B | 0.12 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 36 | Salesforce/blip-itm-large-coco open weights, 0.4B | 0.12 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 37 | laion/CLIP-ViT-B-32-DataComp.XL-s13B-b90K open weights, 0.2B | 0.12 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 38 | laion/CLIP-ViT-B-32-laion2B-s34B-b79K open weights, 0.2B | 0.11 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 39 | jinaai/jina-clip-v1 open weights, 0.2B | 0.11 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 40 | Salesforce/blip-itm-large-flickr open weights, 0.4B | 0.11 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 41 | QuanSun/EVA02-CLIP-B-16 open weights, 0.1B | 0.11 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 42 | Salesforce/blip-itm-base-coco open weights, 0.2B | 0.10 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 43 | openai/clip-vit-large-patch14 open weights, 0.4B | 0.09 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 44 | Salesforce/blip2-opt-2.7b open weights, 3.7B | 0.08 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 45 | openai/clip-vit-base-patch16 open weights, 0.2B | 0.08 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 46 | Salesforce/blip-itm-base-flickr open weights, 0.2B | 0.08 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 47 | Salesforce/blip2-opt-6.7b-coco open weights, 7.8B | 0.08 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 48 | openai/clip-vit-base-patch32 open weights, 0.2B | 0.08 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 49 | nomic-ai/nomic-embed-vision-v1.5 open weights, 0.1B | 0.08 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 50 | BAAI/bge-visualized-base open weights, 0.2B | 0.06 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 51 | Salesforce/blip-image-captioning-large open weights, 0.4B | 0.00 | Text-to-image retrieval | huggingface.co | Official leaderboard | |
| 52 | Salesforce/blip-image-captioning-base open weights, 0.2B | 0.00 | Text-to-image retrieval | huggingface.co | Official leaderboard |
| # | Model | Score | Setting | Date | Link | Source |
|---|---|---|---|---|---|---|
| 1 | | 0.65 | Text-to-image retrieval | Oct 2026 | huggingface.co | Official leaderboard |
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: MTEB’s leaderboard
No lab model card reports it yet
We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.
More in Vision
- MMMU In use 11.5K college-level multimodal questions across 30 subjects. Multiple-choice and open accuracy.
Qwen 3.6 Plus 86%
- MMMU-Pro In use Harder MMMU: college-level multimodal questions with more options and vision-only inputs to remove text shortcuts. GPT-5.6 Sol 84.6%
- CharXiv In use Chart understanding on real arXiv figures. Reasoning questions require synthesizing information from complex charts. Claude Fable 5 93.5%
- MathVista In use Math reasoning in visual contexts (charts, geometry, puzzles) from 31 datasets. Answer accuracy.
Seed2.1 Pro 90.7%
- MATH-Vision (MathVision) In use 3,040 competition math problems with visual context. Answer accuracy.
Qwen 3.8 Max 97.7%
- MMStar In use Models answer 1,500 human-curated image questions that cannot be solved without the image, across 6 core capabilities. Scored by accuracy.
Qwen 3.8 Max 85.9%
Data audit
All 9 checks pass. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 3 pass 2 notes
- Pass Independent runs Scores from the official leaderboard (52 entries).
- Pass Models scored 1 model we track, plus 51 other entries on its leaderboard.
- Pass Newest score The newest is from October 2026, for EmbeddingGemma 2.
- Note Frontier models None of its models are on the capabilities index.
- Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
- Pass Importer We read the board on MTEB: 1 row for models we track and 51 other entries at the last read.
- Pass Last read November 20th, 2026.
- Pass Read worked The last read worked.
Consistency 1 pass 1 note
- Note Lab claims No model has both a lab number and an independent run.
- Pass Known issues No published audit found problems.