Benchmark Almanac

tr_gender_bias

Gemma 4 31B IT 100% leads since June 2026
Not in model cards

Models translate 200 Turkish passages into English. Set B scores pronoun accuracy on 82 passages where a name or kinship term identifies gender.

The author compares 24 translation systems, including GPT 5.5 and Gemini 3.1 Pro. Separate measures track occupational stereotypes and gender choices without a cue.

Released
June 2026
Kind
Translation
Level
Everyday
Scored
Checked automatically
Answers
Open-ended
Tasks
All public
Measures
Propensity
Behavior
Bias
Language
Turkish, English
In and out
Text → Text
Interaction
Single turn
Refresh
One fixed set
Run by
Anyone
Strongest model tested
GPT-5.5 #21 on the capability index
Last model added
May 2026
Scores refreshed
November 20th, 2026
24
Entries in its paper, 13 models we track

3 models share the top score of 100% among models released by May 2026 3 models share the top score of 100% among models released by May 2026

Best score per entry from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. 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. This view shows only models we track, each at its best.

The newest model with a score here came out in May 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 Cohere CAT+ 100%
  2. 2 GoogleTranslate 100%
  3. 3 Lumen 100%
  4. 4 SalamandraTA 100%
  5. 5 TRIVE 100%
  6. 6 VoxNexus_V1 100%
  7. 7 Wayfinder 100%
  8. 8 Gemma 4 31B IT 100%
  9. 9 GPT-5.5 100%
  10. 10 DeepSeek-V4-Pro 100%
  11. 11 Gemma 4-4B 97.6%
  12. 12 Qingqiu-MT-9B 97.6%
  13. 13 Command A+ 95.1%
  14. 14 gpt-oss-120b 93.9%
  15. 15 Qwen3.6 27B 93.9%
  16. 16 Dubformer 92.7%
  17. 17 gpt-oss-20b 91.5%
  18. 18 Mistral Medium 3.5 90.2%
  19. 19 Qwen3.5-9B 87.8%
  20. 20 Tiny Aya Global 85.4%
  21. 21 Tower 9B 84.1%
  22. 22 Ministral 3 14B 84.1%
  23. 23 Gemini 3.1 Pro 81.7%
  24. 24 GLM-4-9B 75.6%

Source: The tr_gender_bias paper

#EntryScoreSettingDateSource
1 Cohere CAT+ 100% Official leaderboard
2 GoogleTranslate 100% Official leaderboard
3 Lumen 100% Official leaderboard
4 SalamandraTA 100% Official leaderboard
5 TRIVE 100% Official leaderboard
6 VoxNexus_V1 100% Official leaderboard
7 Wayfinder 100% Official leaderboard
8 Gemma 4 31B IT 100% Apr 2026 Official leaderboard
9 GPT-5.5 100% Apr 2026 Official leaderboard
10 DeepSeek-V4-Pro 100% Apr 2026 Official leaderboard
11 Gemma 4-4B 97.6% Official leaderboard
12 Qingqiu-MT-9B 97.6% Official leaderboard
13 Command A+ 95.1% May 2026 Official leaderboard
14 gpt-oss-120b 93.9% Aug 2025 Official leaderboard
15 Qwen3.6 27B 93.9% Apr 2026 Official leaderboard
16 Dubformer 92.7% Official leaderboard
17 gpt-oss-20b 91.5% Aug 2025 Official leaderboard
18 Mistral Medium 3.5 90.2% Apr 2026 Official leaderboard
19 Qwen3.5-9B 87.8% Feb 2026 Official leaderboard
20 Tiny Aya Global 85.4% Feb 2026 Official leaderboard
21 Tower 9B 84.1% Official leaderboard
22 Ministral 3 14B 84.1% Dec 2025 Official leaderboard
23 Gemini 3.1 Pro 81.7% Feb 2026 Official leaderboard
24 GLM-4-9B 75.6% Jun 2024 Official leaderboard
#ModelScoreSettingDateSource
1 Gemma 4 31B IT 100% Apr 2026 Official leaderboard
2 GPT-5.5 100% Apr 2026 Official leaderboard
3 DeepSeek-V4-Pro 100% Apr 2026 Official leaderboard
4 Command A+ 95.1% May 2026 Official leaderboard
5 gpt-oss-120b 93.9% Aug 2025 Official leaderboard
6 Qwen3.6 27B 93.9% Apr 2026 Official leaderboard
7 gpt-oss-20b 91.5% Aug 2025 Official leaderboard
8 Mistral Medium 3.5 90.2% Apr 2026 Official leaderboard
9 Qwen3.5-9B 87.8% Feb 2026 Official leaderboard
10 Tiny Aya Global 85.4% Feb 2026 Official leaderboard
11 Ministral 3 14B 84.1% Dec 2025 Official leaderboard
12 Gemini 3.1 Pro 81.7% Feb 2026 Official leaderboard
13 GLM-4-9B 75.6% Jun 2024 Official leaderboard

Rankings by part

We read 12 parts from the benchmark’s paper, each ranking models on one part of the test, such as Malformed, Hedge and Set A they. Gemini 3.1 Pro leads 3 of them, and 6 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.

Malformed

13 models we track have a score here.

  1. 1 Gemini 3.1 Pro ↑11 43%
  2. 2 GLM-4-9B ↑11 0%
  3. 3 gpt-oss-120b ↑2 0%
  4. 4 gpt-oss-20b ↑3 0%
  5. 5 Ministral 3 14B ↑6 0%
  6. 6 Tiny Aya Global ↑4 0%
  7. 7 Qwen3.5-9B ↑2 0%
  8. 8 Gemma 4 31B IT ↓7 0%
  9. 9 Qwen3.6 27B ↓3 0%
  10. 10 GPT-5.5 ↓8 0%

Number one over time

Lab claims count from the day the card came out, independent runs from the model’s release.

Google DeepMind
Benchmark paper Passed 85% in Jun 2026, by Gemma 4 31B IT

Source: The tr_gender_bias paper

Gpt-oss-120b on the official leaderboard passed 85% in August 2025

Models released before it came out already scored that high. All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
In the benchmark’s paper Best so far
0% 20% 40% 60% 80% 100% 2024 2025 2026 Near the ceiling (85%) Gemma 4 31B IT gpt-oss-120b GLM-4-9B 0% 20% 40% 60% 80% 100% 2024 2025 2026 Near the ceiling (85%) Gemma 4 31B IT GLM-4-9B

Source: Official tr_gender_bias 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 Multilingual

All 545 in Multilingual
  • MMMLU In use MMLU professionally translated into 14 languages. Multiple-choice accuracy. Claude Mythos Preview 92.7%
  • WMT24++ In use English-to-55-language translation of WMT24 test data. Scored with automatic metrics such as MetricX or xCOMET. K-EXAONE 90.5%
  • MMLU-ProX In use MMLU-Pro translated into up to 29 languages with expert review. Multiple-choice accuracy per language. Qwen3.7-Max 87%
  • CLIcK In use Models answer 1,995 multiple-choice questions on Korean language and culture. Questions come from official Korean exams and textbooks in eleven categories. Solar Open 2 90.7%
  • Global MMLU Lite 3.0 In use Models answer 400 human-translated MMLU questions per language in 23 languages. Version 3.0 adds Czech, Hungarian, Oriya, Slovak and Tajik and updates Italian. Accuracy. Claude Opus 5 95.7%
  • MMMB In use Models answer multiple-choice questions about images in six languages across 15 categories. The set has about 12,000 questions. LFM2.5-VL-3B 83%

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 (24 entries).
  • Pass Models scored 13 models we track, plus 11 other entries on its leaderboard.
  • Note Newest score The newest came 6 months ago, in May 2026, for Command A+. That’s expected for a saturated test.
  • Note Frontier models None of the top 10 has a score, which is fine for a retired test. The strongest tested is GPT-5.5 (#21).
  • Note Headroom Saturated. The best is 100%, past 85% since August 2025.
Leaderboard import 3 pass
  • Pass Importer We read the board on aclanthology.org: 13 rows for models we track and 11 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