Benchmark Almanac

Finnish Parliament ASR test

Ours: main model (parl) 8.46% tops its leaderboard Whisper large-v3 9.29% leads the models we track
Not in model cards

Models transcribe Finnish parliamentary speeches. The original test reports word error rates separately for speakers heard during training and speakers absent from training.

GAIK scored the 962 segments on the 2016 unseen-speaker list. It could not verify training overlap for every outside model.

Released
August 2017
Kind
Speech recognition and translation
Level
Everyday
Scored
Checked automatically
Answers
Open-ended
Tasks
All public
Measures
Capability
Language
Finnish
In and out
Audio → Text
Interaction
Single turn
Items from
Real work and use
Refresh
One fixed set
Run by
Anyone
In views
Audio and speech
Last model added
October 2026
Scores refreshed
November 20th, 2026
16
Entries on its own leaderboard, 4 models we track

Ours: main model (parl) tops the leaderboard with 8.46% Whisper large-v3 leads with 9.29%

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. The best model we track is Whisper large-v3, at 9.29%. This view shows only models we track, each at its best.

  1. 1 Ours: main model (parl) 8.46%
  2. 2 Ours: timestamps version 8.49%
  3. 3 Finnish-NLP whisper-large-finnish-v3 9.22%
  4. 4 Ours: la-murre version 9.23%
  5. 5 Whisper large-v3 9.29%
  6. 6 Whisper large-v3-turbo 9.69%
  7. 7 NVIDIA Nemotron 3.5 Transcribe 30B-A4B 10.2%
  8. 8 NVIDIA Canary 1B v2 10.2%
  9. 9 Ours: all-skn version 10.6%
  10. 10 Finnish-NLP wav2vec2 XLS-R 1B (no LM) 11.3%
  11. 11 Ours: first model (v1.0) 11.4%
  12. 12 RASMUS Finnish-ASR-Canary-v2, greedy 11.9%
  13. 13 Orukeet (Parakeet v3 fine-tune) 12.7%
  14. 14 NVIDIA Parakeet TDT 0.6B v3 14.6%
  15. 15 RASMUS Finnish-ASR-Canary-v2 + KenLM 16.5%
  16. 16 Qwen3-ASR-1.7B 25.5%

Source: Official Finnish Parliament ASR test leaderboard

Source: Official Finnish Parliament ASR test leaderboard

#EntryScoreSettingDateSource
1 Ours: main model (parl) 8.46% Official leaderboard
2 Ours: timestamps version 8.49% Official leaderboard
3 Finnish-NLP whisper-large-finnish-v3 9.22% Official leaderboard
4 Ours: la-murre version 9.23% Official leaderboard
5 Whisper large-v3 9.29% Nov 2023 Official leaderboard
6 Whisper large-v3-turbo 9.69% Oct 2024 Official leaderboard
7 NVIDIA Nemotron 3.5 Transcribe 30B-A4B 10.2% Oct 2026 Official leaderboard
8 NVIDIA Canary 1B v2 10.2% Official leaderboard
9 Ours: all-skn version 10.6% Official leaderboard
10 Finnish-NLP wav2vec2 XLS-R 1B (no LM) 11.3% Official leaderboard
11 Ours: first model (v1.0) 11.4% Official leaderboard
12 RASMUS Finnish-ASR-Canary-v2, greedy 11.9% Official leaderboard
13 Orukeet (Parakeet v3 fine-tune) 12.7% Official leaderboard
14 NVIDIA Parakeet TDT 0.6B v3 14.6% Official leaderboard
15 RASMUS Finnish-ASR-Canary-v2 + KenLM 16.5% Official leaderboard
16 Qwen3-ASR-1.7B 25.5% Jan 2026 Official leaderboard
#ModelScoreSettingDateSource
1 Whisper large-v3 9.29% Nov 2023 Official leaderboard
2 Whisper large-v3-turbo 9.69% Oct 2024 Official leaderboard
3 NVIDIA Nemotron 3.5 Transcribe 30B-A4B 10.2% Oct 2026 Official leaderboard
4 Qwen3-ASR-1.7B 25.5% Jan 2026 Official leaderboard

Number one over time

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

OpenAI
Own leaderboard 1 leader

Source: Official Finnish Parliament ASR test leaderboard

Scores over time

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
0% 20% 40% 60% 80% 100% 2024 2025 2026 2027 Whisper large-v3 0% 20% 40% 60% 80% 100% 2024 2025 2026 2027 Whisper large-v3

Source: Official Finnish Parliament ASR test 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 Audio and video

All 695 in Audio and video
  • Video-MMMU In use 300 expert-level lecture videos with questions testing perception, comprehension and adaptation of knowledge. Accuracy. Qwen 3.8 Max 88.7%
  • LVBench In use Hours-long video understanding with multiple-choice questions on reasoning, grounding and summarization. Gemini 4 Argon 91.7%
  • MMVU In use 3,000 expert-annotated questions on videos across 27 disciplines. Multiple-choice and open-ended accuracy. Qwen 3.8 Max 82.4%
  • MLVU In use Multi-task long video understanding (minutes to hours) with multiple-choice and generation tasks. Accuracy. Qwen 3.8 Max 90.8%
  • FLEURS In use Models transcribe read speech in 102 languages built from FLoRes-101 sentences. Model cards usually report word error rate, so lower scores are better. Phi-4-multimodal 4.00%
  • VoiceBench In use Spoken instructions with speaker, environment and content variation. Voice assistant responses scored by GPT judge or rules. Qwen3.5-Omni-Plus 93.1%

Data audit

All 9 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 3 pass 2 notes
  • Pass Independent runs Scores from the official leaderboard (16 entries).
  • Pass Models scored 4 models we track, plus 12 other entries on its leaderboard.
  • Pass Newest score The newest is from October 2026, for NVIDIA Nemotron 3.5 Transcribe 30B-A4B.
  • Note Frontier models None of its models are on the capabilities index.
  • Note Headroom Lower is better, so there’s no ceiling to reach.
Leaderboard import 3 pass
  • Pass Importer We read the board on github.com: 4 rows for models we track and 12 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