Benchmark Almanac

Privy Synthetic PII Protocol Trace Dataset

gravitee-small 99.2 tops its leaderboard
Not in model cards

Models locate personal information in synthetic API messages represented as protocol traces. The small and large configurations have separate test files.

Gravitee scores 28,843 small and 120,574 large test examples. PIIMB independently scores a Llama anonymizer on a 5,000-document sample.

Made by
Benjamin Kilimnik and Pixie
Released
September 2022
Kind
Reading comprehension
Level
University
Scored
Checked automatically
Answers
Short answer
Tasks
All public
Measures
Capability
Risk
Privacy
Language
English
In and out
Text, Code → Text
Interaction
Single turn
Items from
Generated
Refresh
One fixed set
Run by
Anyone
In views
Content safety
40
Entries on PII Arena’s leaderboard, 0 models we track

gravitee-small tops the leaderboard with 99.2

Best score per entry from PII Arena’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.

  1. 1 gravitee-small 99.2
  2. 2 pplx 91.8
  3. 3 nym-base 83.4
  4. 4 nym-small 79.1
  5. 5 gliner-nvidia 78.8
  6. 6 gliner2-fastino 77.5
  7. 7 nuner-zero 76.7
  8. 8 gliner2-hivetrace-omni 74.0
  9. 9 apararti 69.3
  10. 10 opf-kz-ru 68.5
  11. 11 gliner-stream-pii 66.1
  12. 12 gliner2-large 64.6
  13. 13 pii-shield-onnx 63.9
  14. 14 openai-base 63.8
  15. 15 gliner25-fastino 61.2
  16. 16 mmbert32k 60.4
  17. 17 opf-ru 59.9
  18. 18 gliner-urchade 59.3
  19. 19 gliner-pii-base 57.6
  20. 20 gliner2-hivetrace-uni 56.8
  21. 21 kalyan-ettin 55.5
  22. 22 opf-ru-v2 51.1
  23. 23 traciora 49.8
  24. 24 ru-pii-ner 49.3
  25. 25 gliner-pii-edge 48.5
  26. 26 openmed-nemotron 42.1
  27. 27 gliner-multi-v21 39.2
  28. 28 ru-legal-ner 35.2
  29. 29 bardsai-eu 35.0
  30. 30 gliner2-vladlinv 33.3
  31. 31 davlan-mbert 30.3
  32. 32 davlan-xlmr 26.5
  33. 33 ner-ru-yqelz 23.5
  34. 34 ner-ru-gherman 20.6
  35. 35 rules-ru 15.2
  36. 36 stanza-ru 11.1
  37. 37 natasha 8.85
  38. 38 fef2-secret-ru 6.42
  39. 39 spacy-ru-lg 3.21
  40. 40 spacy-alrosait 0.00

Source: PII Arena’s leaderboard

#EntryScoreSettingDateLinkSource
1 gravitee-small 99.2 piiarena.com Official leaderboard
2 pplx 91.8 piiarena.com Official leaderboard
3 nym-base 83.4 piiarena.com Official leaderboard
4 nym-small 79.1 piiarena.com Official leaderboard
5 gliner-nvidia 78.8 piiarena.com Official leaderboard
6 gliner2-fastino 77.5 piiarena.com Official leaderboard
7 nuner-zero 76.7 piiarena.com Official leaderboard
8 gliner2-hivetrace-omni 74.0 piiarena.com Official leaderboard
9 apararti 69.3 piiarena.com Official leaderboard
10 opf-kz-ru 68.5 piiarena.com Official leaderboard
11 gliner-stream-pii 66.1 piiarena.com Official leaderboard
12 gliner2-large 64.6 piiarena.com Official leaderboard
13 pii-shield-onnx 63.9 piiarena.com Official leaderboard
14 openai-base 63.8 piiarena.com Official leaderboard
15 gliner25-fastino 61.2 piiarena.com Official leaderboard
16 mmbert32k 60.4 piiarena.com Official leaderboard
17 opf-ru 59.9 piiarena.com Official leaderboard
18 gliner-urchade 59.3 piiarena.com Official leaderboard
19 gliner-pii-base 57.6 piiarena.com Official leaderboard
20 gliner2-hivetrace-uni 56.8 piiarena.com Official leaderboard
21 kalyan-ettin 55.5 piiarena.com Official leaderboard
22 opf-ru-v2 51.1 piiarena.com Official leaderboard
23 traciora 49.8 piiarena.com Official leaderboard
24 ru-pii-ner 49.3 piiarena.com Official leaderboard
25 gliner-pii-edge 48.5 piiarena.com Official leaderboard
26 openmed-nemotron 42.1 piiarena.com Official leaderboard
27 gliner-multi-v21 39.2 piiarena.com Official leaderboard
28 ru-legal-ner 35.2 piiarena.com Official leaderboard
29 bardsai-eu 35.0 piiarena.com Official leaderboard
30 gliner2-vladlinv 33.3 piiarena.com Official leaderboard
31 davlan-mbert 30.3 piiarena.com Official leaderboard
32 davlan-xlmr 26.5 piiarena.com Official leaderboard
33 ner-ru-yqelz 23.5 piiarena.com Official leaderboard
34 ner-ru-gherman 20.6 piiarena.com Official leaderboard
35 rules-ru 15.2 piiarena.com Official leaderboard
36 stanza-ru 11.1 piiarena.com Official leaderboard
37 natasha 8.85 piiarena.com Official leaderboard
38 fef2-secret-ru 6.42 piiarena.com Official leaderboard
39 spacy-ru-lg 3.21 piiarena.com Official leaderboard
40 spacy-alrosait 0.00 piiarena.com 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 Knowledge

All 332 in Knowledge
  • MMLU-Pro In use Harder 10-option version of MMLU with reasoning-focused questions across 14 subjects. Intern-S2-397B 89.8%
  • MMLU In use 57-subject multiple-choice exam questions from elementary to professional level. Scored by accuracy. Seed1.8 92.3%
  • SuperGPQA In use 26,529 graduate-level multiple-choice questions across 285 disciplines. Accuracy. Qwen 3.6 Max (Preview) 73.9%
  • WideSearch In use 200 bilingual broad info-seeking tasks: agent collects many atomic facts into a table. Scored by success rate and row/item F1. Tencent Hy4 preview 83.9%
  • BBEH micro average In use Models solve harder replacements for each of the 23 BIG-Bench Hard tasks, 4,520 examples in total. The score is the share of all examples answered correctly. Gemma 4 31B IT 74.4%
  • HLE-Diamond In use Models answer 1,000 cleaned expert questions from Humanity's Last Exam, 500 on reasoning and 500 on knowledge, closed book. Graded on exact answers. Claude Fable 5.2 77.9%

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 official leaderboard (40 entries).
  • Pass Models scored 40 entries on its leaderboard, none a model we track.
  • Note Newest score No scores yet.
  • Note Frontier models No models scored yet.
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on PII Arena: 40 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