Benchmark Almanac

Voice Arena Goal Dataset

Not in model cards

Judges compare two calls and choose which caller achieved the airline task more successfully. Human preferences provide the reference.

Voice Arena publishes the 200 conversations and pairwise judgments. The group also runs MonsoonASR and public speech synthesis leaderboards.

Released
September 2026
Kind
Audio understanding and voice
Level
Everyday
Scored
Ranked by votes
Answers
Multiple choice
Tasks
All public
Measures
Preference
Language
English
In and out
Audio, Text → Audio
Interaction
Multi-agent
Environment
The live world
Items from
Live play or votes
Refresh
One fixed set
Run by
A host
Last model added
March 2025
Scores refreshed
November 21st, 2026
3
Entries in its paper, 2 models we track

Qwen2.5-Omni-7B leads with 64.7% Qwen2.5-Omni-7B leads with 64.7%

Best score per entry from the benchmark's paper, read on November 21st, 2026. A paper’s table doesn’t add newer models. The paper reports Task and Humanness accuracy, so this headline uses the Task column. 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.

  1. 1 Qwen2.5-Omni-7B text 64.7%
  2. 2 Phi-4-multimodal audio 49.5%
  3. 3 Qwen3-Omni-30B audio 38.4%

Source: The Voice Arena Goal Dataset paper

Source: The Voice Arena Goal Dataset paper

#EntryScoreSettingDateSource
1 Qwen2.5-Omni-7B 64.7% text Mar 2025 Official leaderboard
2 Phi-4-multimodal 49.5% audio Feb 2025 Official leaderboard
3 Qwen3-Omni-30B 38.4% audio Official leaderboard
#ModelScoreSettingDateSource
1 Qwen2.5-Omni-7B 64.7% text Mar 2025 Official leaderboard
2 Phi-4-multimodal 49.5% audio Feb 2025 Official leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

In the benchmark’s paper Best so far
0% 20% 40% 60% 80% 100% 2025 Near the ceiling (85%) Qwen2.5-Omni-7B Phi-4-multimodal 0% 20% 40% 60% 80% 100% 2025 Near the ceiling (85%) Qwen2.5-Omni-7B Phi-4-multimodal

Source: Official Voice Arena Goal Dataset 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

8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 20 months ago, in March 2025, for Qwen2.5-Omni-7B. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 2 pass 2 notes
  • Pass Independent runs Scores from the benchmark’s paper (3 entries).
  • Note Models scored Only 2 models we track, plus 1 other entry on its leaderboard.
  • Warning Newest score The newest came 20 months ago, in March 2025, for Qwen2.5-Omni-7B.
  • Note Frontier models None of its models are on the capabilities index.
  • Pass Headroom The best is 64.7%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on huggingface.co: 2 rows for models we track and 1 other entry at the last read.
  • Pass Last read November 21st, 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