Benchmark Almanac

People Making Conversation

Moshi 0.39 leads since October 2026
Not in model cards

Speech models continue real conversations at recorded pauses and turn endings. The benchmark compares their timing, response length, and recall with the original speakers.

Made by
Arya Bastani and colleagues
Released
October 2026
Kind
Speech recognition and translation
Level
Everyday
Scored
Checked automatically
Answers
Agent task
Tasks
All public
Measures
Capability
Language
English
In and out
Audio → Audio
Interaction
Single turn
Items from
Real work and use
Refresh
One fixed set
Run by
Anyone
In views
Audio and speech
Last model added
March 2026
Scores refreshed
November 20th, 2026
5
Entries in its paper, 3 models we track

Moshi leads with 0.39 among models released by March 2026 Moshi leads with 0.39 among models released by March 2026

Best score per entry from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. The paper reports no overall score. The headline is the fraction of pause points at which the system stayed silent. 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 March 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 Moshi 0.39
  2. 2 PersonaPlex 0.22
  3. 3 gpt-realtime semantic, low 0.05
  4. 4 Grok Voice 0.00
  5. 5 Gemini 3.1 Flash Live 0.00

Source: The People Making Conversation paper

Source: The People Making Conversation paper

#EntryScore ±SettingDateSource
1 Moshi 0.39 ±0.1 Sep 2024 Official leaderboard
2 PersonaPlex 0.22 0.2 to 0.3 Official leaderboard
3 gpt-realtime 0.05 0 to 0.1 semantic, low Aug 2025 Official leaderboard
4 Grok Voice 0.00 Official leaderboard
5 Gemini 3.1 Flash Live 0.00 Mar 2026 Official leaderboard
#ModelScore ±SettingDateSource
1 Moshi 0.39 ±0.1 Sep 2024 Official leaderboard
2 gpt-realtime 0.05 0 to 0.1 semantic, low Aug 2025 Official leaderboard
3 Gemini 3.1 Flash Live 0.00 Mar 2026 Official leaderboard

Rankings by part

We read 8 parts from the benchmark’s paper, each ranking models on one part of the test, such as Pause takeover, After pause and Turn-end latency. gpt-realtime leads 6 of them, and Moshi leads the other 2. Arrows show how far a model moved from its overall rank among the models we track.

Pause takeover

3 models we track have a score here.

  1. 1 gpt-realtime ↑1 0.06
  2. 2 Gemini 3.1 Flash Live ↑1 0.18
  3. 3 Moshi ↓2 0.29

Number one over time

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

Kyutai
Benchmark paper Not tested since Oct 2026
Moshi 0.39

Source: The People Making Conversation paper

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
In the benchmark’s paper
0.00 0.10 0.20 0.30 0.40 2025 2026 0.00 0.10 0.20 0.30 0.40 2025 2026

Source: Official People Making Conversation 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 8 months ago, in March 2026, for Gemini 3.1 Flash Live. 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 (5 entries).
  • Pass Models scored 3 models we track, plus 2 other entries on its leaderboard.
  • Warning Newest score The newest came 8 months ago, in March 2026, for Gemini 3.1 Flash Live.
  • Note Frontier models None of its models are on the capabilities index.
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on arxiv.org: 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