Benchmark Almanac

NWPU-RESISC45

QuanSun/EVA02-CLIP-bigE-14-plus 92.8% tops its leaderboard EmbeddingGemma 2 88.3% leads the models we track
Not in model cards

Models classify remote sensing images into 45 scene categories.

MTEB reports accuracy on its RESISC45 task.

Released
March 2017
Kind
General visual questions
Level
University
Scored
Checked automatically
Answers
Prediction
Tasks
All public
Measures
Capability
Language
English
In and out
Images → Text
Interaction
Single turn
Items from
Real work and use
Refresh
One fixed set
Run by
Anyone
In views
Vision
Last model added
October 2026
Scores refreshed
November 20th, 2026
57
Entries on MTEB’s leaderboard, 1 model we track

QuanSun/EVA02-CLIP-bigE-14-plus tops the leaderboard with 92.8% EmbeddingGemma 2 leads with 88.3%

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 88.3%. This view shows only models we track, each at its best.

57 entries. Scroll sideways for the rest.

  1. 92.8 QuanSun/EVA02-CLIP-bigE-14-plus
  2. 92.0 QuanSun/EVA02-CLIP-bigE-14
  3. 91.6 google/siglip-so400m-patch14-384
  4. 90.9 google/siglip-large-patch16-384
  5. 90.9 laion/CLIP-ViT-bigG-14-laion2B-39B-b160k
  6. 90.9 laion/CLIP-ViT-H-14-laion2B-s32B-b79K
  7. 90.7 laion/CLIP-ViT-g-14-laion2B-s34B-b88K
  8. 90.4 laion/CLIP-ViT-L-14-DataComp.XL-s13B-b90K
  9. 90.4 jinaai/jina-embeddings-v5-omni-small
  10. 90.3 BidirLM/BidirLM-Omni-2.5B-Embedding
  11. 90.1 laion/CLIP-ViT-L-14-laion2B-s32B-b82K
  12. 89.9 google/siglip-large-patch16-256
  13. 89.6 openai/clip-vit-large-patch14
  14. 89.1 Haon-Chen/e5-omni-7B
  15. 89.1 royokong/e5-v
  16. 89.1 LCO-Embedding/LCO-Embedding-Omni-3B
  17. 88.9 LCO-Embedding/LCO-Embedding-Omni-7B
  18. 88.6 dnotitia/DNA-VL-STEER-2B
  19. 88.6 Salesforce/blip2-opt-2.7b
  20. 88.4 google/siglip-base-patch16-384
  21. 88.4 google/siglip-base-patch16-512
  22. 88.3 EmbeddingGemma 2
  23. 88.2 Tevatron/OmniEmbed-v0.1
  24. 88.1 encord-team/ebind-full
  25. 87.7 google/siglip-base-patch16-224
  26. 87.6 laion/CLIP-ViT-B-16-DataComp.XL-s13B-b90K
  27. 87.6 Salesforce/blip2-opt-6.7b-coco
  28. 87.5 google/siglip-base-patch16-256
  29. 87.3 mongodb/voyage-multimodal-3
  30. 86.5 openai/clip-vit-base-patch16
  31. 86.5 Haon-Chen/e5-omni-3B
  32. 86.3 facebook/dinov2-giant
  33. 86.2 facebook/dinov2-large
  34. 85.8 laion/CLIP-ViT-B-32-laion2B-s34B-b79K
  35. 85.7 Salesforce/blip-itm-large-coco
  36. 85.6 google/siglip-base-patch16-256-multilingual
  37. 84.8 laion/CLIP-ViT-B-32-DataComp.XL-s13B-b90K
  38. 84.5 QuanSun/EVA02-CLIP-L-14
  39. 84.0 jinaai/jina-clip-v1
  40. 83.8 kakaobrain/align-base
  41. 83.1 BAAI/bge-visualized-base
  42. 83.0 Salesforce/blip-itm-large-flickr
  43. 82.8 openai/clip-vit-base-patch32
  44. 82.0 Salesforce/blip-image-captioning-large
  45. 81.9 Salesforce/blip-itm-base-coco
  46. 81.3 facebook/dinov2-base
  47. 81.2 jinaai/jina-embeddings-v5-omni-nano
  48. 81.0 BAAI/bge-visualized-m3
  49. 79.4 TIGER-Lab/VLM2Vec-Full
  50. 79.4 TIGER-Lab/VLM2Vec-LoRA
  51. 78.5 Salesforce/blip-itm-base-flickr
  52. 77.9 nomic-ai/nomic-embed-vision-v1.5
  53. 76.9 facebook/dinov2-small
  54. 76.2 QuanSun/EVA02-CLIP-B-16
  55. 75.1 Salesforce/blip-image-captioning-base
  56. 73.8 nyu-visionx/moco-v3-vit-l
  57. 72.5 nyu-visionx/moco-v3-vit-b
  1. 1 QuanSun/EVA02-CLIP-bigE-14-plus 92.8%
  2. 2 QuanSun/EVA02-CLIP-bigE-14 92%
  3. 3 google/siglip-so400m-patch14-384 91.6%
  4. 4 google/siglip-large-patch16-384 90.9%
  5. 5 laion/CLIP-ViT-bigG-14-laion2B-39B-b160k 90.9%
  6. 6 laion/CLIP-ViT-H-14-laion2B-s32B-b79K 90.9%
  7. 7 laion/CLIP-ViT-g-14-laion2B-s34B-b88K 90.7%
  8. 8 laion/CLIP-ViT-L-14-DataComp.XL-s13B-b90K 90.4%
  9. 9 jinaai/jina-embeddings-v5-omni-small 90.4%
  10. 10 BidirLM/BidirLM-Omni-2.5B-Embedding 90.3%
  11. 11 laion/CLIP-ViT-L-14-laion2B-s32B-b82K 90.1%
  12. 12 google/siglip-large-patch16-256 89.9%
  13. 13 openai/clip-vit-large-patch14 89.6%
  14. 14 Haon-Chen/e5-omni-7B 89.1%
  15. 15 royokong/e5-v 89.1%
  16. 16 LCO-Embedding/LCO-Embedding-Omni-3B 89.1%
  17. 17 LCO-Embedding/LCO-Embedding-Omni-7B 88.9%
  18. 18 dnotitia/DNA-VL-STEER-2B 88.6%
  19. 19 Salesforce/blip2-opt-2.7b 88.6%
  20. 20 google/siglip-base-patch16-384 88.4%
  21. 21 google/siglip-base-patch16-512 88.4%
  22. 22 EmbeddingGemma 2 88.3%
  23. 23 Tevatron/OmniEmbed-v0.1 88.2%
  24. 24 encord-team/ebind-full 88.1%
  25. 25 google/siglip-base-patch16-224 87.7%
  26. 26 laion/CLIP-ViT-B-16-DataComp.XL-s13B-b90K 87.6%
  27. 27 Salesforce/blip2-opt-6.7b-coco 87.6%
  28. 28 google/siglip-base-patch16-256 87.5%
  29. 29 mongodb/voyage-multimodal-3 87.3%
  30. 30 openai/clip-vit-base-patch16 86.5%
  31. 31 Haon-Chen/e5-omni-3B 86.5%
  32. 32 facebook/dinov2-giant 86.3%
  33. 33 facebook/dinov2-large 86.2%
  34. 34 laion/CLIP-ViT-B-32-laion2B-s34B-b79K 85.8%
  35. 35 Salesforce/blip-itm-large-coco 85.7%
  36. 36 google/siglip-base-patch16-256-multilingual 85.6%
  37. 37 laion/CLIP-ViT-B-32-DataComp.XL-s13B-b90K 84.8%
  38. 38 QuanSun/EVA02-CLIP-L-14 84.5%
  39. 39 jinaai/jina-clip-v1 84%
  40. 40 kakaobrain/align-base 83.8%
  41. 41 BAAI/bge-visualized-base 83.1%
  42. 42 Salesforce/blip-itm-large-flickr 83%
  43. 43 openai/clip-vit-base-patch32 82.8%
  44. 44 Salesforce/blip-image-captioning-large 82%
  45. 45 Salesforce/blip-itm-base-coco 81.9%
  46. 46 facebook/dinov2-base 81.3%
  47. 47 jinaai/jina-embeddings-v5-omni-nano 81.2%
  48. 48 BAAI/bge-visualized-m3 81%
  49. 49 TIGER-Lab/VLM2Vec-Full 79.4%
  50. 50 TIGER-Lab/VLM2Vec-LoRA 79.4%
  51. 51 Salesforce/blip-itm-base-flickr 78.5%
  52. 52 nomic-ai/nomic-embed-vision-v1.5 77.9%
  53. 53 facebook/dinov2-small 76.9%
  54. 54 QuanSun/EVA02-CLIP-B-16 76.2%
  55. 55 Salesforce/blip-image-captioning-base 75.1%
  56. 56 nyu-visionx/moco-v3-vit-l 73.8%
  57. 57 nyu-visionx/moco-v3-vit-b 72.5%

Source: MTEB’s leaderboard

Source: MTEB’s leaderboard

#EntryScoreSettingDateSource
1 QuanSun/EVA02-CLIP-bigE-14-plus open weights, 5B 92.8% Official leaderboard
2 QuanSun/EVA02-CLIP-bigE-14 open weights, 4.7B 92% Official leaderboard
3 google/siglip-so400m-patch14-384 open weights, 0.9B 91.6% Official leaderboard
4 google/siglip-large-patch16-384 open weights, 0.7B 90.9% Official leaderboard
5 laion/CLIP-ViT-bigG-14-laion2B-39B-b160k open weights, 2.5B 90.9% Official leaderboard
6 laion/CLIP-ViT-H-14-laion2B-s32B-b79K open weights, 1B 90.9% Official leaderboard
7 laion/CLIP-ViT-g-14-laion2B-s34B-b88K open weights, 1.4B 90.7% Official leaderboard
8 laion/CLIP-ViT-L-14-DataComp.XL-s13B-b90K open weights, 0.4B 90.4% Official leaderboard
9 jinaai/jina-embeddings-v5-omni-small open weights, 1.6B 90.4% Official leaderboard
10 BidirLM/BidirLM-Omni-2.5B-Embedding open weights, 2.4B 90.3% Official leaderboard
11 laion/CLIP-ViT-L-14-laion2B-s32B-b82K open weights, 0.4B 90.1% Official leaderboard
12 google/siglip-large-patch16-256 open weights, 0.7B 89.9% Official leaderboard
13 openai/clip-vit-large-patch14 open weights, 0.4B 89.6% Official leaderboard
14 Haon-Chen/e5-omni-7B open weights, 8.9B 89.1% Official leaderboard
15 royokong/e5-v open weights, 8.4B 89.1% Official leaderboard
16 LCO-Embedding/LCO-Embedding-Omni-3B open weights, 4.7B 89.1% Official leaderboard
17 LCO-Embedding/LCO-Embedding-Omni-7B open weights, 8.9B 88.9% Official leaderboard
18 dnotitia/DNA-VL-STEER-2B open weights, 2.1B 88.6% Official leaderboard
19 Salesforce/blip2-opt-2.7b open weights, 3.7B 88.6% Official leaderboard
20 google/siglip-base-patch16-384 open weights, 0.2B 88.4% Official leaderboard
21 google/siglip-base-patch16-512 open weights, 0.2B 88.4% Official leaderboard
22 EmbeddingGemma 2 88.3% Oct 2026 Official leaderboard
23 Tevatron/OmniEmbed-v0.1 open weights, 8.9B 88.2% Official leaderboard
24 encord-team/ebind-full open weights, 1.8B 88.1% Official leaderboard
25 google/siglip-base-patch16-224 open weights, 0.2B 87.7% Official leaderboard
26 laion/CLIP-ViT-B-16-DataComp.XL-s13B-b90K open weights, 0.2B 87.6% Official leaderboard
27 Salesforce/blip2-opt-6.7b-coco open weights, 7.8B 87.6% Official leaderboard
28 google/siglip-base-patch16-256 open weights, 0.2B 87.5% Official leaderboard
29 mongodb/voyage-multimodal-3 closed weights 87.3% Official leaderboard
30 openai/clip-vit-base-patch16 open weights, 0.2B 86.5% Official leaderboard
31 Haon-Chen/e5-omni-3B open weights, 4.7B 86.5% Official leaderboard
32 facebook/dinov2-giant open weights, 1.1B 86.3% Official leaderboard
33 facebook/dinov2-large open weights, 0.3B 86.2% Official leaderboard
34 laion/CLIP-ViT-B-32-laion2B-s34B-b79K open weights, 0.2B 85.8% Official leaderboard
35 Salesforce/blip-itm-large-coco open weights, 0.4B 85.7% Official leaderboard
36 google/siglip-base-patch16-256-multilingual open weights, 0.4B 85.6% Official leaderboard
37 laion/CLIP-ViT-B-32-DataComp.XL-s13B-b90K open weights, 0.2B 84.8% Official leaderboard
38 QuanSun/EVA02-CLIP-L-14 open weights, 0.4B 84.5% Official leaderboard
39 jinaai/jina-clip-v1 open weights, 0.2B 84% Official leaderboard
40 kakaobrain/align-base open weights, 0.2B 83.8% Official leaderboard
41 BAAI/bge-visualized-base open weights, 0.2B 83.1% Official leaderboard
42 Salesforce/blip-itm-large-flickr open weights, 0.4B 83% Official leaderboard
43 openai/clip-vit-base-patch32 open weights, 0.2B 82.8% Official leaderboard
44 Salesforce/blip-image-captioning-large open weights, 0.4B 82% Official leaderboard
45 Salesforce/blip-itm-base-coco open weights, 0.2B 81.9% Official leaderboard
46 facebook/dinov2-base open weights, 0.1B 81.3% Official leaderboard
47 jinaai/jina-embeddings-v5-omni-nano open weights, 1B 81.2% Official leaderboard
48 BAAI/bge-visualized-m3 open weights, 0.9B 81% Official leaderboard
49 TIGER-Lab/VLM2Vec-Full open weights, 4.1B 79.4% Official leaderboard
50 TIGER-Lab/VLM2Vec-LoRA open weights, 4.2B 79.4% Official leaderboard
51 Salesforce/blip-itm-base-flickr open weights, 0.2B 78.5% Official leaderboard
52 nomic-ai/nomic-embed-vision-v1.5 open weights, 0.1B 77.9% Official leaderboard
53 facebook/dinov2-small open weights, 0B 76.9% Official leaderboard
54 QuanSun/EVA02-CLIP-B-16 open weights, 0.1B 76.2% Official leaderboard
55 Salesforce/blip-image-captioning-base open weights, 0.2B 75.1% Official leaderboard
56 nyu-visionx/moco-v3-vit-l open weights, 0.3B 73.8% Official leaderboard
57 nyu-visionx/moco-v3-vit-b open weights, 0.1B 72.5% Official leaderboard
#ModelScoreSettingDateSource
1 EmbeddingGemma 2 88.3% Oct 2026 Official leaderboard

EmbeddingGemma 2 on the official leaderboard passed 85% in October 2026

That was 115 months after its release. All the scores we have, one dot each, so a model tested at several settings appears more than once.

On MTEB’s leaderboard
0% 20% 40% 60% 80% 100% Near the ceiling (85%) EmbeddingGemma 2 0% 20% 40% 60% 80% 100% Near the ceiling (85%) EmbeddingGemma 2

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

All 636 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 (57 entries).
  • Pass Models scored 1 model we track, plus 56 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 Saturated. The best is 88.3%, past 85% since October 2026.
Leaderboard import 3 pass
  • Pass Importer We read the board on MTEB: 1 row for models we track and 56 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.

How we check