Swedish Benchmarks V1.1
Version 1.1
Models answer Swedish language, document, and knowledge questions. The overall score combines normalized task performance and response time.
Markovic added three knowledge tests on April 29, 2026. The August leaderboard ranks 23 models on all 13 tests.
- Made by
- Drasko Markovic
- Released
- April 2026
- Kind
- Reading in many languages
- Level
- School
- Scored
- Checked automatically
- Answers
- Short answer
- Measures
- Capability
- Language
- Swedish
- Interaction
- Single turn
- Items from
- Another benchmark
- Refresh
- New versions replace it
- Run by
- A host
- Based on
- Swedish Benchmarks V1.0
- In views
- Law and finance
- Strongest model tested
-
Qwen 3.8 27B #59 on the capability index
- Last model added
- August 2026
- Scores refreshed
- November 20th, 2026
Gemma 4 31B IT leads with 85.5 among models released by August 2026 Gemma 4 31B IT leads with 85.5 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:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 16 more
- 85.5
Gemma 4 31B IT
- 80.7
Qwen3.6 27B
- 80.2
Gemma 4 26B A4B
- 79.1 qwen3.5-122b-a10b-halo
- 78.4 gpt-oss-120b
- 78.3 qwen3.6-27b-mtp
- 77.9
Qwen 3.6 35B-A3B
- 77.4 nvidia_nemotron-3-super
- 73.8
Qwen 3.8 27B
- 70.8
MiniMax-M2.7
- 70.2
Llama 3.3 70B
- 69.7
Gemma 3 27B
- 68.6 nvidia_nemotron-3-nano
- 66.0 gpt-oss-20b
- 63.2 nvidia_nemotron-3-nano-omni_thinking
- 63.2
Muse Glimmer
- 62.9 zai-org_glm-4.7-flash
- 59.5
Nemotron 3.5 Lightning 30B-A3B
- 57.3 poolside_laguna-s-2.1
- 56.5 nvidia_nemotron-3-nano-omni
- 52.1 mistralai_ministral-3-14b-reasoning
- 44.4 mistral-small-4-119b-2603
- 44.2 Granite 4.2 30B
- 1
Gemma 4 31B IT 85.5
- 2
Qwen3.6 27B 80.7
- 3
Gemma 4 26B A4B 80.2
- 4 qwen3.5-122b-a10b-halo 79.1
- 5 gpt-oss-120b 78.4
- 6 qwen3.6-27b-mtp 78.3
- 7
Qwen 3.6 35B-A3B 77.9
- 8 nvidia_nemotron-3-super 77.4
- 9
Qwen 3.8 27B 73.8
- 10
MiniMax-M2.7 70.8
- 11
Llama 3.3 70B 70.2
- 12
Gemma 3 27B 69.7
- 13 nvidia_nemotron-3-nano 68.6
- 14 gpt-oss-20b 66.0
- 15 nvidia_nemotron-3-nano-omni_thinking 63.2
- 16
Muse Glimmer 63.2
- 17 zai-org_glm-4.7-flash 62.9
- 18
Nemotron 3.5 Lightning 30B-A3B 59.5
- 19 poolside_laguna-s-2.1 57.3
- 20 nvidia_nemotron-3-nano-omni 56.5
- 21 mistralai_ministral-3-14b-reasoning 52.1
- 22 mistral-small-4-119b-2603 44.4
- 23 Granite 4.2 30B 44.2
Source: Official Swedish Benchmarks V1.1 leaderboard
- 1
Gemma 4 31B IT 85.5
- 2
Qwen3.6 27B 80.7
- 3
Gemma 4 26B A4B 80.2
- 4 gpt-oss-120b 78.4
- 5
Qwen 3.6 35B-A3B 77.9
- 6
Qwen 3.8 27B 73.8
- 7
MiniMax-M2.7 70.8
- 8
Llama 3.3 70B 70.2
- 9
Gemma 3 27B 69.7
- 10 gpt-oss-20b 66.0
- 11
Muse Glimmer 63.2
- 12
Nemotron 3.5 Lightning 30B-A3B 59.5
- 13 Granite 4.2 30B 44.2
Source: Official Swedish Benchmarks V1.1 leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 85.5 | Apr 2026 | Official leaderboard | |
| 2 | | 80.7 | Apr 2026 | Official leaderboard | |
| 3 | | 80.2 | Apr 2026 | Official leaderboard | |
| 4 | qwen3.5-122b-a10b-halo | 79.1 | Official leaderboard | ||
| 5 | gpt-oss-120b | 78.4 | Aug 2025 | Official leaderboard | |
| 6 | qwen3.6-27b-mtp | 78.3 | Official leaderboard | ||
| 7 | | 77.9 | Apr 2026 | Official leaderboard | |
| 8 | nvidia_nemotron-3-super | 77.4 | Official leaderboard | ||
| 9 | | 73.8 | Aug 2026 | Official leaderboard | |
| 10 | | 70.8 | Mar 2026 | Official leaderboard | |
| 11 | | 70.2 | Dec 2024 | Official leaderboard | |
| 12 | | 69.7 | Mar 2025 | Official leaderboard | |
| 13 | nvidia_nemotron-3-nano | 68.6 | Official leaderboard | ||
| 14 | gpt-oss-20b | 66.0 | Aug 2025 | Official leaderboard | |
| 15 | nvidia_nemotron-3-nano-omni_thinking | 63.2 | Official leaderboard | ||
| 16 | | 63.2 | Aug 2026 | Official leaderboard | |
| 17 | zai-org_glm-4.7-flash | 62.9 | Official leaderboard | ||
| 18 | | 59.5 | Aug 2026 | Official leaderboard | |
| 19 | poolside_laguna-s-2.1 | 57.3 | Official leaderboard | ||
| 20 | nvidia_nemotron-3-nano-omni | 56.5 | Official leaderboard | ||
| 21 | mistralai_ministral-3-14b-reasoning | 52.1 | Official leaderboard | ||
| 22 | mistral-small-4-119b-2603 | 44.4 | Official leaderboard | ||
| 23 | Granite 4.2 30B | 44.2 | Aug 2026 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 85.5 | Apr 2026 | Official leaderboard | |
| 2 | | 80.7 | Apr 2026 | Official leaderboard | |
| 3 | | 80.2 | Apr 2026 | Official leaderboard | |
| 4 | gpt-oss-120b | 78.4 | Aug 2025 | Official leaderboard | |
| 5 | | 77.9 | Apr 2026 | Official leaderboard | |
| 6 | | 73.8 | Aug 2026 | Official leaderboard | |
| 7 | | 70.8 | Mar 2026 | Official leaderboard | |
| 8 | | 70.2 | Dec 2024 | Official leaderboard | |
| 9 | | 69.7 | Mar 2025 | Official leaderboard | |
| 10 | gpt-oss-20b | 66.0 | Aug 2025 | Official leaderboard | |
| 11 | | 63.2 | Aug 2026 | Official leaderboard | |
| 12 | | 59.5 | Aug 2026 | Official leaderboard | |
| 13 | Granite 4.2 30B | 44.2 | Aug 2026 | Official leaderboard |
Rankings by part
We read 13 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as New Synonym Benchmark, SveaKnow Culture and SveaKnow Economy. Gemma 4 31B IT leads 6 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.
New Synonym Benchmark
13 models we track have a score here.
- 1
Gemma 4 31B IT 99.3%
- 2
Gemma 4 26B A4B ↑1 95.1%
- 3
Gemma 3 27B ↑6 91.3%
- 4
Qwen 3.6 35B-A3B ↑1 89.2%
- 5
Qwen3.6 27B ↓3 89.1%
- 6
Llama 3.3 70B ↑2 87.2%
- 7
Qwen 3.8 27B ↓1 82.7%
- 8 gpt-oss-120b ↓4 79.4%
- 9 gpt-oss-20b ↑1 76.4%
- 10
Nemotron 3.5 Lightning 30B-A3B ↑2 76%
SveaKnow Culture
13 models we track have a score here.
- 1
Llama 3.3 70B ↑7 95.4%
- 2
Qwen 3.6 35B-A3B ↑3 89.3%
- 3
Gemma 4 31B IT ↓2 87.5%
- 4
Qwen3.6 27B ↓2 87.4%
- 5
Gemma 4 26B A4B ↓2 86.6%
- 6 gpt-oss-120b ↓2 83.5%
- 7
Gemma 3 27B ↑2 82.8%
- 8
MiniMax-M2.7 ↓1 81.6%
- 9
Qwen 3.8 27B ↓3 71.6%
- 10
Nemotron 3.5 Lightning 30B-A3B ↑2 67.2%
SveaKnow Economy
13 models we track have a score here.
- 1
Qwen3.6 27B ↑1 93.6%
- 2
Gemma 3 27B ↑7 91.1%
- 3 gpt-oss-120b ↑1 89.6%
- 4
Gemma 4 31B IT ↓3 89.6%
- 5
Qwen 3.6 35B-A3B 81%
- 6
Llama 3.3 70B ↑2 80.1%
- 7
Gemma 4 26B A4B ↓4 79.3%
- 8
MiniMax-M2.7 ↓1 77.8%
- 9
Nemotron 3.5 Lightning 30B-A3B ↑3 67.6%
- 10
Qwen 3.8 27B ↓4 63.6%
SveaKnow Society
13 models we track have a score here.
- 1
Gemma 4 31B IT 97.9%
- 2
Gemma 4 26B A4B ↑1 97.5%
- 3 gpt-oss-120b ↑1 90.6%
- 4
Llama 3.3 70B ↑4 87.2%
- 5
MiniMax-M2.7 ↑2 81.6%
- 6
Qwen3.6 27B ↓4 79.1%
- 7
Gemma 3 27B ↑2 77.6%
- 8
Qwen 3.6 35B-A3B ↓3 75.2%
- 9
Muse Glimmer ↑2 71.3%
- 10
Qwen 3.8 27B ↓4 63.4%
SweDiagnostics
13 models we track have a score here.
- 1
Gemma 4 31B IT 81.9%
- 2 gpt-oss-120b ↑2 80.1%
- 3
Qwen 3.6 35B-A3B ↑2 76%
- 4
Qwen3.6 27B ↓2 65%
- 5 gpt-oss-20b ↑5 64.9%
- 6
Gemma 3 27B ↑3 62.1%
- 7
Qwen 3.8 27B ↓1 61.8%
- 8
MiniMax-M2.7 ↓1 58.2%
- 9
Gemma 4 26B A4B ↓6 56.1%
- 10
Muse Glimmer ↑1 54.9%
SweFAQ
13 models we track have a score here.
- 1
Qwen3.6 27B ↑1 89.2%
- 2
Qwen 3.6 35B-A3B ↑3 88%
- 3 gpt-oss-120b ↑1 87.9%
- 4
Qwen 3.8 27B ↑2 84.3%
- 5
Muse Glimmer ↑6 80.3%
- 6
Gemma 4 26B A4B ↓3 74.8%
- 7
Gemma 4 31B IT ↓6 72.6%
- 8 gpt-oss-20b ↑2 68.3%
- 9
Llama 3.3 70B ↓1 66.9%
- 10 Granite 4.2 30B ↑3 63.1%
SweNLI
13 models we track have a score here.
- 1
Gemma 4 31B IT 91.6%
- 2 gpt-oss-120b ↑2 89.4%
- 3
Qwen 3.8 27B ↑3 74.6%
- 4 gpt-oss-20b ↑6 74.1%
- 5
Muse Glimmer ↑6 68.2%
- 6
Gemma 4 26B A4B ↓3 68.2%
- 7
MiniMax-M2.7 66.1%
- 8
Qwen 3.6 35B-A3B ↓3 65.7%
- 9
Qwen3.6 27B ↓7 64.2%
- 10 Granite 4.2 30B ↑3 60%
SweParaphrase
13 models we track have a score here.
- 1
Gemma 4 31B IT 97.9%
- 2
Gemma 4 26B A4B ↑1 97.8%
- 3
Qwen 3.6 35B-A3B ↑2 96.9%
- 4
Qwen3.6 27B ↓2 95.3%
- 5
Gemma 3 27B ↑4 92.7%
- 6
Qwen 3.8 27B 91.9%
- 7 gpt-oss-20b ↑3 90.5%
- 8
Llama 3.3 70B 90.1%
- 9
Nemotron 3.5 Lightning 30B-A3B ↑3 87.2%
- 10 Granite 4.2 30B ↑3 86%
SwePDF
13 models we track have a score here.
- 1
Qwen 3.6 35B-A3B ↑4 93.2%
- 2
Qwen3.6 27B 87.3%
- 3
Qwen 3.8 27B ↑3 85.6%
- 4
MiniMax-M2.7 ↑3 84.7%
- 5 gpt-oss-120b ↓1 84.1%
- 6 gpt-oss-20b ↑4 82.8%
- 7
Gemma 3 27B ↑2 70.8%
- 8
Muse Glimmer ↑3 70.2%
- 9
Llama 3.3 70B ↓1 69.9%
- 10
Gemma 4 26B A4B ↓7 69.6%
SweSAT Synonyms
13 models we track have a score here.
- 1
Gemma 3 27B ↑8 91.5%
- 2
Gemma 4 31B IT ↓1 91.3%
- 3
Llama 3.3 70B ↑5 88.1%
- 4
Qwen 3.6 35B-A3B ↑1 87.2%
- 5
Gemma 4 26B A4B ↓2 82.2%
- 6
MiniMax-M2.7 ↑1 78.7%
- 7
Muse Glimmer ↑4 75%
- 8
Qwen3.6 27B ↓6 73.5%
- 9
Qwen 3.8 27B ↓3 62.8%
- 10 gpt-oss-120b ↓6 49.7%
SweWiC
13 models we track have a score here.
- 1
Gemma 4 26B A4B ↑2 95.5%
- 2
Qwen 3.8 27B ↑4 94%
- 3
Qwen3.6 27B ↓1 93.3%
- 4
Gemma 4 31B IT ↓3 93.3%
- 5 gpt-oss-120b ↓1 85.4%
- 6
MiniMax-M2.7 ↑1 82.3%
- 7 Granite 4.2 30B ↑6 79.6%
- 8
Muse Glimmer ↑3 70.5%
- 9
Qwen 3.6 35B-A3B ↓4 70.3%
- 10
Nemotron 3.5 Lightning 30B-A3B ↑2 64.2%
SweWinoGender
13 models we track have a score here.
- 1 gpt-oss-20b ↑9 63.3%
- 2
Gemma 4 26B A4B ↑1 53.3%
- 3 gpt-oss-120b ↑1 50.3%
- 4
Gemma 4 31B IT ↓3 49.8%
- 5
Qwen3.6 27B ↓3 41.5%
- 6
Llama 3.3 70B ↑2 36.8%
- 7
Gemma 3 27B ↑2 35.9%
- 8
Nemotron 3.5 Lightning 30B-A3B ↑4 35.8%
- 9
Qwen 3.8 27B ↓3 34.8%
- 10
Qwen 3.6 35B-A3B ↓5 30.8%
SweWinograd
13 models we track have a score here.
- 1
Gemma 4 31B IT 93.1%
- 2
Qwen3.6 27B 90.8%
- 3
Qwen 3.8 27B ↑3 87.8%
- 4
Gemma 4 26B A4B ↓1 87%
- 5
Llama 3.3 70B ↑3 82.1%
- 6
Muse Glimmer ↑5 80.5%
- 7
MiniMax-M2.7 77%
- 8
Nemotron 3.5 Lightning 30B-A3B ↑4 74.2%
- 9
Qwen 3.6 35B-A3B ↓4 69.7%
- 10 gpt-oss-20b 68.1%
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.
Source: Official Swedish Benchmarks V1.1 leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official Swedish Benchmarks V1.1 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.
Versions
More 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. The strongest tested is Qwen 3.8 27B (#59).
- Note Headroom Scored as a plain number, with no fixed ceiling.
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.