Benchmark Almanac

Revisited Oxford and Paris

HesAff–rSIFT–ASMK⋆ +SP → R– (ref. 10)–R-MAC+DFS 74.7% tops its leaderboard
Not in model cards

Models retrieve photographs of the same landmark from Oxford and Paris image collections. The revised benchmark defines medium and hard evaluation protocols.

Released
March 2018
Kind
General visual questions
Level
Everyday
Scored
Checked automatically
Answers
Short answer
Tasks
All public
Measures
Capability
Language
English
In and out
Images → Text
Interaction
Single turn
Run by
Anyone
In views
Vision
31
Entries in its paper, 0 models we track

HesAff–rSIFT–ASMK⋆ +SP → R– (ref. 10)–R-MAC+DFS tops the leaderboard with 74.7%

Best score per entry from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. The paper reports no overall score, so this is the plain mean of its eight mean average precision scores. 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.

  1. 1 HesAff–rSIFT–ASMK⋆ +SP → R– (ref. 10)–R-MAC+DFS 74.7%
  2. 2 HesAff–rSIFT–ASMK⋆ +SP → R– (ref. 37)–GeM+DFS 73.2%
  3. 3 DELF–ASMK⋆ +SP → R– (ref. 10)–R-MAC+DFS 70.5%
  4. 4 R – (ref. 37) – GeM+DFS 66.1%
  5. 5 R – (ref. 10) – R-MAC+DFS 65.2%
  6. 6 V – (ref. 37) – GeM+DFS 63.7%
  7. 7 DELF–HQE+SP 59.6%
  8. 8 R – (ref. 37) – GeM+αQE 51.6%
  9. 9 DELF–ASMK⋆ +SP 51.5%
  10. 10 R – (ref. 10) – R-MAC+αQE 51.4%
  11. 11 R – (ref. 37) – GeM 47.4%
  12. 12 V – (ref. 37) – GeM+αQE 47.1%
  13. 13 R – (ref. 10) – R-MAC 45.8%
  14. 14 V – (ref. 37) – GeM 41.9%
  15. 15 R – [O] – R-MAC+αQE 40.9%
  16. 16 V – (ref. 36) – MAC 39.4%
  17. 17 R – [O] – R-MAC 37.3%
  18. 18 R – [O] – GeM 34.9%
  19. 19 R – [O] – MAC 32.4%
  20. 20 R – [O] – CroW 32.1%
  21. 21 R – [O] – SPoC 30.9%
  22. 22 V – [O] – GeM 30.4%
  23. 23 V – [O] – R-MAC 30%
  24. 24 V – [O] – CroW 28.1%
  25. 25 V – [O] – MAC 27.9%
  26. 26 A – (ref. 37) – GeM 27.5%
  27. 27 V – (ref. 1) – NetVLAD 27%
  28. 28 A – (ref. 36) – MAC 26.8%
  29. 29 V – [O] – SPoC 24.7%
  30. 30 A – [O] – GeM 21.6%
  31. 31 A – [O] – MAC 18.5%

Source: The Revisited Oxford and Paris paper

#EntryScoreSettingDateSource
1 HesAff–rSIFT–ASMK⋆ +SP → R– (ref. 10)–R-MAC+DFS 74.7% Official leaderboard
2 HesAff–rSIFT–ASMK⋆ +SP → R– (ref. 37)–GeM+DFS 73.2% Official leaderboard
3 DELF–ASMK⋆ +SP → R– (ref. 10)–R-MAC+DFS 70.5% Official leaderboard
4 R – (ref. 37) – GeM+DFS 66.1% Official leaderboard
5 R – (ref. 10) – R-MAC+DFS 65.2% Official leaderboard
6 V – (ref. 37) – GeM+DFS 63.7% Official leaderboard
7 DELF–HQE+SP 59.6% Official leaderboard
8 R – (ref. 37) – GeM+αQE 51.6% Official leaderboard
9 DELF–ASMK⋆ +SP 51.5% Official leaderboard
10 R – (ref. 10) – R-MAC+αQE 51.4% Official leaderboard
11 R – (ref. 37) – GeM 47.4% Official leaderboard
12 V – (ref. 37) – GeM+αQE 47.1% Official leaderboard
13 R – (ref. 10) – R-MAC 45.8% Official leaderboard
14 V – (ref. 37) – GeM 41.9% Official leaderboard
15 R – [O] – R-MAC+αQE 40.9% Official leaderboard
16 V – (ref. 36) – MAC 39.4% Official leaderboard
17 R – [O] – R-MAC 37.3% Official leaderboard
18 R – [O] – GeM 34.9% Official leaderboard
19 R – [O] – MAC 32.4% Official leaderboard
20 R – [O] – CroW 32.1% Official leaderboard
21 R – [O] – SPoC 30.9% Official leaderboard
22 V – [O] – GeM 30.4% Official leaderboard
23 V – [O] – R-MAC 30% Official leaderboard
24 V – [O] – CroW 28.1% Official leaderboard
25 V – [O] – MAC 27.9% Official leaderboard
26 A – (ref. 37) – GeM 27.5% Official leaderboard
27 V – (ref. 1) – NetVLAD 27% Official leaderboard
28 A – (ref. 36) – MAC 26.8% Official leaderboard
29 V – [O] – SPoC 24.7% Official leaderboard
30 A – [O] – GeM 21.6% Official leaderboard
31 A – [O] – MAC 18.5% Official 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 8 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 2 pass 3 notes
  • Pass Independent runs Scores from the benchmark’s paper (31 entries).
  • Pass Models scored 31 entries on its leaderboard, none a model we track.
  • Note Newest score No scores yet.
  • Note Frontier models No models scored yet.
  • Note Headroom No scores yet.
Leaderboard import 3 pass
  • Pass Importer We read the board on openaccess.thecvf.com: 31 entries at the last read, none a model we track.
  • 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