Benchmark Almanac

Jarvis Bench

Human Agent 1,295 tops its leaderboard Gemini 3.8 Live 998 leads the models we track
Not in model cards

Voice agents handle conversational tasks. Listeners rate naturalness, and the board reports task effectiveness separately.

The published rows retain the domain, system name, and votes.

Made by
Voice Arena
Kind
Speech recognition and translation
Level
Everyday
Scored
Ranked by votes
Answers
Open-ended
Measures
Preference
In and out
Audio → Audio
Interaction
Multi-turn
Refresh
One fixed set
Run by
A host
In views
Audio and speech
Last model added
September 2026
Scores refreshed
November 20th, 2026
7
Entries on its own leaderboard, 5 models we track

Human Agent tops the leaderboard with 1,295 Gemini 3.8 Live leads with 998 among models released by September 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. The best model we track is Gemini 3.8 Live, at 998. This view shows only models we track, each at its best.

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

Bars start at 900.

  1. 1 Human Agent 1,295
  2. 2 Gemini 3.8 Live 998
  3. 3 GPT-Realtime-2.1 995
  4. 4 Gemini 3.1 Flash Live 981
  5. 5 Gemini 3.8 Live Extended Thinking high 963
  6. 6 GPT-Live-1 (GPT 6 Astra, Medium) 951
  7. 7 Grok Voice 2.0 915

Source: Official Jarvis Bench leaderboard

#EntryScore ±SettingDateSource
1 Human Agent 1,295 1263 to 1334 Official leaderboard
2 Gemini 3.8 Live 998 985 to 1010 Sep 2026 Official leaderboard
3 GPT-Realtime-2.1 995 ±16 Jul 2026 Official leaderboard
4 Gemini 3.1 Flash Live 981 965 to 998 Mar 2026 Official leaderboard
5 Gemini 3.8 Live Extended Thinking 963 ±16 high Sep 2026 Official leaderboard
6 GPT-Live-1 (GPT 6 Astra, Medium) 951 934 to 969 Official leaderboard
7 Grok Voice 2.0 915 ±19 Jul 2026 Official leaderboard
#ModelScore ±SettingDateSource
1 Gemini 3.8 Live 998 985 to 1010 Sep 2026 Official leaderboard
2 GPT-Realtime-2.1 995 ±16 Jul 2026 Official leaderboard
3 Gemini 3.1 Flash Live 981 965 to 998 Mar 2026 Official leaderboard
4 Gemini 3.8 Live Extended Thinking 963 ±16 high Sep 2026 Official leaderboard
5 Grok Voice 2.0 915 ±19 Jul 2026 Official leaderboard

Rankings by part

We read 2 parts from the benchmark’s own leaderboard: General conversations and Airlines. Each ranks models on one part of the test. Gemini 3.8 Live leads both. Arrows show how far a model moved from its overall rank among the models we track.

General conversations

5 models we track have a score here.

  1. 1 Gemini 3.8 Live 1,002
  2. 2 GPT-Realtime-2.1 999
  3. 3 Gemini 3.1 Flash Live 983
  4. 4 Gemini 3.8 Live Extended Thinking 952
  5. 5 Grok Voice 2.0 923

3 models have held number one on the official leaderboard since 2023

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

Google DeepMind OpenAI
Own leaderboard Not tested since Sep 2026

Source: Official Jarvis Bench 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
900 920 940 960 980 1,000 2026 900 920 940 960 980 1,000 2026

Source: Official Jarvis Bench 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 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 (7 entries).
  • Pass Models scored 5 models we track, plus 2 other entries on its leaderboard.
  • Pass Newest score The newest is from September 2026, for Gemini 3.8 Live.
  • Note Frontier models None of its models are on the capabilities index.
  • Note Headroom Scored as a rating, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on voicearena.com: 5 rows for models we track and 2 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