Kiji PII Detection
Models locate and label personal information in multilingual text. Dataiku publishes a test split, and independent evaluators compare predicted spans with its annotations.
Uppala evaluated GPT-4o on 846 English items. Gravitee evaluated 1,033. The February and May English test sets share 265 distinct texts.
- Made by
- Dataiku
- Released
- February 2026
- Kind
- Following precise instructions
- Level
- Everyday
- Scored
- Checked automatically
- Answers
- Short answer
- Tasks
- All public
- Measures
- Capability
- Risk
- Privacy
- Language
- Many languages
- Interaction
- Single turn
- Refresh
- New versions replace it
- Run by
- Anyone
- In views
- Many languages, Content safety
nym-small tops the leaderboard with 96.1
Best score per entry from the benchmark's own 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.
- 96.1 nym-small
- 96.0 gliner2-fastino
- 95.7 pplx
- 95.4 nym-base
- 95.3 gravitee-small
- 95.0 gliner2-hivetrace-omni
- 93.3 openmed-multilingual
- 93.1 gliner-nvidia
- 89.7 bardsai-eu
- 89.1 openmed-nemotron
- 86.0 kalyan-ettin
- 85.9 gliner2-large
- 85.8 gliner-stream-pii
- 84.8 apararti
- 84.4 openai-base
- 82.8 opf-kz-ru
- 77.5 mmbert32k
- 76.6 traciora
- 74.1 gliner25-fastino
- 73.7 nuner-zero
- 73.6 opf-ru-v2
- 71.4 gliner2-hivetrace-uni
- 70.4 stanza-ru
- 69.2 opf-ru
- 67.3 gliner-multi-v21
- 66.8 ru-pii-ner
- 62.1 gliner-pii-base
- 60.1 gliner-urchade
- 60.0 davlan-xlmr
- 59.1 gliner-pii-edge
- 56.3 davlan-mbert
- 56.3 ner-ru-gherman
- 54.9 gliner2-vladlinv
- 51.6 ner-ru-yqelz
- 37.6 pii-shield-onnx
- 28.0 ru-legal-ner
- 20.5 spacy-ru-lg
- 14.0 natasha
- 13.7 fef2-secret-ru
- 7.84 rules-ru
- 0.00 spacy-alrosait
- 1 nym-small 96.1
- 2 gliner2-fastino 96.0
- 3 pplx 95.7
- 4 nym-base 95.4
- 5 gravitee-small 95.3
- 6 gliner2-hivetrace-omni 95.0
- 7 openmed-multilingual 93.3
- 8 gliner-nvidia 93.1
- 9 bardsai-eu 89.7
- 10 openmed-nemotron 89.1
- 11 kalyan-ettin 86.0
- 12 gliner2-large 85.9
- 13 gliner-stream-pii 85.8
- 14 apararti 84.8
- 15 openai-base 84.4
- 16 opf-kz-ru 82.8
- 17 mmbert32k 77.5
- 18 traciora 76.6
- 19 gliner25-fastino 74.1
- 20 nuner-zero 73.7
- 21 opf-ru-v2 73.6
- 22 gliner2-hivetrace-uni 71.4
- 23 stanza-ru 70.4
- 24 opf-ru 69.2
- 25 gliner-multi-v21 67.3
- 26 ru-pii-ner 66.8
- 27 gliner-pii-base 62.1
- 28 gliner-urchade 60.1
- 29 davlan-xlmr 60.0
- 30 gliner-pii-edge 59.1
- 31 davlan-mbert 56.3
- 32 ner-ru-gherman 56.3
- 33 gliner2-vladlinv 54.9
- 34 ner-ru-yqelz 51.6
- 35 pii-shield-onnx 37.6
- 36 ru-legal-ner 28.0
- 37 spacy-ru-lg 20.5
- 38 natasha 14.0
- 39 fef2-secret-ru 13.7
- 40 rules-ru 7.84
- 41 spacy-alrosait 0.00
Source: Official Kiji PII Detection leaderboard
| # | Entry | Score | Setting | Date | Link | Source |
|---|---|---|---|---|---|---|
| 1 | nym-small | 96.1 | piiarena.com | Official leaderboard | ||
| 2 | gliner2-fastino | 96.0 | piiarena.com | Official leaderboard | ||
| 3 | pplx | 95.7 | piiarena.com | Official leaderboard | ||
| 4 | nym-base | 95.4 | piiarena.com | Official leaderboard | ||
| 5 | gravitee-small | 95.3 | piiarena.com | Official leaderboard | ||
| 6 | gliner2-hivetrace-omni | 95.0 | piiarena.com | Official leaderboard | ||
| 7 | openmed-multilingual | 93.3 | piiarena.com | Official leaderboard | ||
| 8 | gliner-nvidia | 93.1 | piiarena.com | Official leaderboard | ||
| 9 | bardsai-eu | 89.7 | piiarena.com | Official leaderboard | ||
| 10 | openmed-nemotron | 89.1 | piiarena.com | Official leaderboard | ||
| 11 | kalyan-ettin | 86.0 | piiarena.com | Official leaderboard | ||
| 12 | gliner2-large | 85.9 | piiarena.com | Official leaderboard | ||
| 13 | gliner-stream-pii | 85.8 | piiarena.com | Official leaderboard | ||
| 14 | apararti | 84.8 | piiarena.com | Official leaderboard | ||
| 15 | openai-base | 84.4 | piiarena.com | Official leaderboard | ||
| 16 | opf-kz-ru | 82.8 | piiarena.com | Official leaderboard | ||
| 17 | mmbert32k | 77.5 | piiarena.com | Official leaderboard | ||
| 18 | traciora | 76.6 | piiarena.com | Official leaderboard | ||
| 19 | gliner25-fastino | 74.1 | piiarena.com | Official leaderboard | ||
| 20 | nuner-zero | 73.7 | piiarena.com | Official leaderboard | ||
| 21 | opf-ru-v2 | 73.6 | piiarena.com | Official leaderboard | ||
| 22 | gliner2-hivetrace-uni | 71.4 | piiarena.com | Official leaderboard | ||
| 23 | stanza-ru | 70.4 | piiarena.com | Official leaderboard | ||
| 24 | opf-ru | 69.2 | piiarena.com | Official leaderboard | ||
| 25 | gliner-multi-v21 | 67.3 | piiarena.com | Official leaderboard | ||
| 26 | ru-pii-ner | 66.8 | piiarena.com | Official leaderboard | ||
| 27 | gliner-pii-base | 62.1 | piiarena.com | Official leaderboard | ||
| 28 | gliner-urchade | 60.1 | piiarena.com | Official leaderboard | ||
| 29 | davlan-xlmr | 60.0 | piiarena.com | Official leaderboard | ||
| 30 | gliner-pii-edge | 59.1 | piiarena.com | Official leaderboard | ||
| 31 | davlan-mbert | 56.3 | piiarena.com | Official leaderboard | ||
| 32 | ner-ru-gherman | 56.3 | piiarena.com | Official leaderboard | ||
| 33 | gliner2-vladlinv | 54.9 | piiarena.com | Official leaderboard | ||
| 34 | ner-ru-yqelz | 51.6 | piiarena.com | Official leaderboard | ||
| 35 | pii-shield-onnx | 37.6 | piiarena.com | Official leaderboard | ||
| 36 | ru-legal-ner | 28.0 | piiarena.com | Official leaderboard | ||
| 37 | spacy-ru-lg | 20.5 | piiarena.com | Official leaderboard | ||
| 38 | natasha | 14.0 | piiarena.com | Official leaderboard | ||
| 39 | fef2-secret-ru | 13.7 | piiarena.com | Official leaderboard | ||
| 40 | rules-ru | 7.84 | piiarena.com | Official leaderboard | ||
| 41 | 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 Instructions and chat
- IFEval In use Models answer about 500 prompts containing verifiable instructions such as word limits or required keywords. Code checks each instruction for strict and loose accuracy.
Gemma 4 31B IT 98.9%
- IFBench In use 58 new out-of-domain verifiable output constraints. Scored by programmatic constraint checks (strict or loose). Flint 89.2%
- MultiChallenge In use Multi-turn conversations testing instruction retention, inference memory, versioned editing and self-coherence. Graded by LLM judge with rubrics. GPT-6.1 Sol 82%
- Multi-IF In use Models follow verifiable instructions over three-turn conversations in eight languages, extending IFEval. Scored by instruction accuracy at each turn. LFM3-1.7B 89.1%
- AdvancedIF In use Models respond to over 1,600 prompts with complex, multi-turn, and system-level instructions. A judge model checks each response against expert-written rubrics.
MAI-Thinking-1 85%
- MM-IFEval In use Multimodal models follow image-grounded instructions with exact output constraints. Checks mix rule-based verification and a judge model. GPT-4o 64.6%
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 (41 entries).
- Pass Models scored 41 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 piiarena.com: 41 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.