Benchmark Almanac

New Synonym Benchmark

Gemma 4 31B IT 99.3% leads since April 2026
Not in model cards

Models choose the one Swedish word among five options that does not mean the same thing as the target word.

Markovic’s leaderboard reports accuracy for 23 models on 150 questions.

Made by
Drasko Markovic
Released
April 2026
Kind
Reading in many languages
Level
Everyday
Scored
Checked automatically
Answers
Multiple choice
Measures
Capability
Language
Swedish
In and out
Text → Text
Interaction
Single turn
Refresh
One fixed set
Run by
A host
Strongest model tested
Qwen 3.8 27B #59 on the capability index
Last model added
August 2026
Scores refreshed
November 20th, 2026
23
Entries on its own leaderboard, 13 models we track

Gemma 4 31B IT leads with 99.3% among models released by August 2026 Gemma 4 31B IT leads with 99.3% among models released by August 2026

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

The newest model with a score here came out in August 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 Gemma 4 31B IT 99.3%
  2. 2 Gemma 4 26B A4B 95.2%
  3. 3 qwen3.5-122b-a10b-halo 94%
  4. 4 Gemma 3 27B 91.3%
  5. 5 Qwen 3.6 35B-A3B 89.2%
  6. 6 Qwen3.6 27B 89.1%
  7. 7 qwen3.6-27b-mtp 88.8%
  8. 8 nvidia_nemotron-3-super 87.3%
  9. 9 Llama 3.3 70B 87.2%
  10. 10 Qwen 3.8 27B 82.7%
  11. 11 gpt-oss-120b 79.4%
  12. 12 gpt-oss-20b 76.4%
  13. 13 Nemotron 3.5 Lightning 30B-A3B 76%
  14. 14 MiniMax-M2.7 74%
  15. 15 Muse Glimmer 73.1%
  16. 16 nvidia_nemotron-3-nano 70.1%
  17. 17 nvidia_nemotron-3-nano-omni 67%
  18. 18 zai-org_glm-4.7-flash 61.5%
  19. 19 nvidia_nemotron-3-nano-omni_thinking 61.1%
  20. 20 poolside_laguna-s-2.1 52.9%
  21. 21 Granite 4.2 30B 48.5%
  22. 22 mistralai_ministral-3-14b-reasoning 32.7%
  23. 23 mistral-small-4-119b-2603 24.2%

Source: Official New Synonym Benchmark leaderboard

#EntryScoreSettingDateSource
1 Gemma 4 31B IT 99.3% Apr 2026 Official leaderboard
2 Gemma 4 26B A4B 95.2% Apr 2026 Official leaderboard
3 qwen3.5-122b-a10b-halo 94% Official leaderboard
4 Gemma 3 27B 91.3% Mar 2025 Official leaderboard
5 Qwen 3.6 35B-A3B 89.2% Apr 2026 Official leaderboard
6 Qwen3.6 27B 89.1% Apr 2026 Official leaderboard
7 qwen3.6-27b-mtp 88.8% Official leaderboard
8 nvidia_nemotron-3-super 87.3% Official leaderboard
9 Llama 3.3 70B 87.2% Dec 2024 Official leaderboard
10 Qwen 3.8 27B 82.7% Aug 2026 Official leaderboard
11 gpt-oss-120b 79.4% Aug 2025 Official leaderboard
12 gpt-oss-20b 76.4% Aug 2025 Official leaderboard
13 Nemotron 3.5 Lightning 30B-A3B 76% Aug 2026 Official leaderboard
14 MiniMax-M2.7 74% Mar 2026 Official leaderboard
15 Muse Glimmer 73.1% Aug 2026 Official leaderboard
16 nvidia_nemotron-3-nano 70.1% Official leaderboard
17 nvidia_nemotron-3-nano-omni 67% Official leaderboard
18 zai-org_glm-4.7-flash 61.5% Official leaderboard
19 nvidia_nemotron-3-nano-omni_thinking 61.1% Official leaderboard
20 poolside_laguna-s-2.1 52.9% Official leaderboard
21 Granite 4.2 30B 48.5% Aug 2026 Official leaderboard
22 mistralai_ministral-3-14b-reasoning 32.7% Official leaderboard
23 mistral-small-4-119b-2603 24.2% Official leaderboard
#ModelScoreSettingDateSource
1 Gemma 4 31B IT 99.3% Apr 2026 Official leaderboard
2 Gemma 4 26B A4B 95.2% Apr 2026 Official leaderboard
3 Gemma 3 27B 91.3% Mar 2025 Official leaderboard
4 Qwen 3.6 35B-A3B 89.2% Apr 2026 Official leaderboard
5 Qwen3.6 27B 89.1% Apr 2026 Official leaderboard
6 Llama 3.3 70B 87.2% Dec 2024 Official leaderboard
7 Qwen 3.8 27B 82.7% Aug 2026 Official leaderboard
8 gpt-oss-120b 79.4% Aug 2025 Official leaderboard
9 gpt-oss-20b 76.4% Aug 2025 Official leaderboard
10 Nemotron 3.5 Lightning 30B-A3B 76% Aug 2026 Official leaderboard
11 MiniMax-M2.7 74% Mar 2026 Official leaderboard
12 Muse Glimmer 73.1% Aug 2026 Official leaderboard
13 Granite 4.2 30B 48.5% Aug 2026 Official leaderboard

2 models have held number one on the official leaderboard since its release in April 2026

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

Google DeepMind
Own leaderboard Passed 85% in Apr 2026, by Gemma 3 27B

Source: Official New Synonym Benchmark leaderboard

Llama 3.3 70B on the official leaderboard passed 85% in December 2024

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
On the benchmark’s own leaderboard Best so far
0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) Gemma 4 31B IT Gemma 3 27B 0% 20% 40% 60% 80% 100% 2025 2026 Near the ceiling (85%) Gemma 4 31B IT Gemma 3 27B

Source: Official New Synonym Benchmark 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

8 of 9 checks pass, 1 warning. Look first at the Task access check. We don’t know whether the tasks are public. Every benchmark

Catalog record 1 warning 1 pass
  • Warning Task access We don’t know whether the tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 3 pass 2 notes
  • Pass Independent runs Scores from the official leaderboard (23 entries).
  • Pass Models scored 13 models we track, plus 10 other entries on its leaderboard.
  • Pass Newest score The newest is from August 2026, for Granite 4.2 30B.
  • Note Frontier models None of the top 10 has a score, which is fine for a retired test. The strongest tested is Qwen 3.8 27B (#59).
  • Note Headroom Saturated. The best is 99.3%, past 85% since December 2024.
Leaderboard import 3 pass
  • Pass Importer We read the board on markovic.com: 13 rows for models we track and 10 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