DeepgramAudio modelActive
Flux
PopularityRank 319 of 876
37out of 100same as last month
Summary
Deepgram built Flux for voice agents that need to listen, think, and recognize the right moment to respond. The streaming model combines speech recognition with model-integrated end-of-turn detection and configurable turn-taking. The flux-general-en variant handles English across accents and returns turn events alongside the transcript.
Flux is limited to English interactive agents with model-integrated turn taking. For multilingual or traditional batch transcription, Nova-3 remains the broader Deepgram path.
All benchmark results
Every published result with its source, test setup and comparison models.
Show all 234 published benchmark results
Page 1 of 20
Measurements without matching peer values
These measurements have no matching peer values under the same test conditions. Their original values and sources remain available here.
VoiceCodeBench: CTEM (Show measurement, test conditions, and source)
- Source value
- 86.572
- Score
- 86.572
- Metric
- canonical token/entity match
- Unit
- %
- Benchmark version
- 1
- Category
- general
- Task
- ctem
- Task label
- CTEM
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- Vals AI
- Status
- active
- Retrieved at
- 2026-11-06
- Methodology
- Source model ID: deepgram/flux-general-en. Source model name: Flux. Provider reported by source: Deepgram. Evaluator: Vals AI. Dataset type: public. Leaderboard date: 2026-09-24. Latency in seconds: 68.8. Standard error: 0. Cost per task in USD: 0.007172. Source ID: vals-ai. Evidence capture ID: vals-voice-code-bench-2026-11-06-7753d162bfe3. Source payload hash: 7753d162bfe3fa3470d586da3917bf8ddd96f14e10f9aa8f03dd517b8d5f003b. Parser version: vals-astro-v4. Imported at: 2026-11-06. Source cohort: vals-voice-code-bench:1:ctem
VoiceCodeBench: Overall (Show measurement, test conditions, and source)
- Source value
- 55.667
- Score
- 55.667
- Metric
- task success rate
- Unit
- %
- Benchmark version
- 1
- Category
- general
- Task
- overall
- Task label
- Overall
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- Vals AI
- Status
- active
- Retrieved at
- 2026-11-06
- Methodology
- Source model ID: deepgram/flux-general-en. Source model name: Flux. Provider reported by source: Deepgram. Evaluator: Vals AI. Dataset type: public. Leaderboard date: 2026-09-24. Latency in seconds: 68.8. Standard error: 0. Cost per task in USD: 0.007172. Source ID: vals-ai. Evidence capture ID: vals-voice-code-bench-2026-11-06-7753d162bfe3. Source payload hash: 7753d162bfe3fa3470d586da3917bf8ddd96f14e10f9aa8f03dd517b8d5f003b. Parser version: vals-astro-v4. Imported at: 2026-11-06. Source cohort: vals-voice-code-bench:1:overall
Artificial Analysis AA-WER Streaming (final transcription) (Show measurement, test conditions, and source)
- Source value
- 7.3903
- Score
- 7.3903
- Metric
- word error rate (WER)
- Unit
- %
- Benchmark version
- AA-WER v2.2 streaming
- Category
- audio
- Direction
- Lower is better
- Source type
- independent-evaluator
- Evaluator
- Artificial Analysis
- Status
- active
- Retrieved at
- 2026-10-15
- Methodology
- Independent evaluation. Direct API streaming evaluation; final-transcription WER across AA-AgentTalk (50%), VoxPopuli (25%), and Earnings22 (25%), about eight hours total. WER is duration-weighted within dataset; Earnings22 chunking is reassembled at parent-call level. AA-WER v2.2 streaming protocol.
Coval pooled corpus-level STT word error rate (Show measurement, test conditions, and source)
- Source value
- 5.789713965267468
- Score
- 5.789713965267468
- Metric
- Pooled corpus-level speech-recognition word error rate, Coval pooled_value
- Unit
- %
- Benchmark version
- Coval rolling 30-day window, re-measured daily; route last measured 2026-10-16
- Category
- audio
- Direction
- Lower is better
- Sample
- 26027
- Source type
- official-leaderboard
- Evaluator
- Coval
- Status
- superseded
- Retrieved at
- 2026-10-16
- Methodology
- Official benchmark leaderboard. Coval endpoint key deepgram/flux-general-en, exact API model ID flux-general-en, hosted by Deepgram. The route identifies the English streaming recognizer separately from Flux General Multi. Rolling 30-day window, re-measured daily.
- Context
- This is Coval's pooled corpus-level STT WER, not the separate per-clip mean_value or a rounded leaderboard headline. Pooled WER and mean per-clip WER are different aggregations and should not be compared as the same score. The WER row's sample_count is used, not another metric's count or a roster total. The result describes Coval's hosted 30-day endpoint cohort, not local or native performance.
Cekura Converse-STT, Deepgram Flux English Pipecat WER (Show measurement, test conditions, and source)
- Source value
- 3.9
- Score
- 3.9
- Metric
- word error rate (WER) on Pipecat public short-clip speech-to-text set
- Unit
- %
- Benchmark version
- Cekura Converse-STT page last updated 2026-09-16, retrieved 2026-10-17
- Category
- audio
- Direction
- Lower is better
- Sample
- 999
- Lower confidence bound
- 3.53
- Upper confidence bound
- 4.29
- Source type
- independent-evaluator
- Evaluator
- Cekura Bench
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Independent evaluation. Streaming STT row named Deepgram Flux English, matched to the official API model ID flux-general-en on /v2/listen. Pipecat has 1,000 public clips; the row reports 999 usable model outputs. Cekura labels the uncertainty estimate a likely 95% range. The endpoint build is not pinned.
- Context
- The published interval is Cekura's likely 95% range, not a separately documented model confidence interval. The 999 valid outputs, rather than all 1,000 corpus clips, are the sampleCount. This distinct public-clip WER is not the existing Coval, AA-WER, or VoiceCodeBench result.
Cekura Converse-STT, Deepgram Flux English Ocular WER (Show measurement, test conditions, and source)
- Source value
- 4.18
- Score
- 4.18
- Metric
- word error rate (WER) on Ocular licensed real-phone-conversation speech-to-text set
- Unit
- %
- Benchmark version
- Cekura Converse-STT page last updated 2026-09-16, retrieved 2026-10-17
- Category
- audio
- Direction
- Lower is better
- Sample
- 8
- Lower confidence bound
- 3.08
- Upper confidence bound
- 5.23
- Source type
- independent-evaluator
- Evaluator
- Cekura Bench
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Independent evaluation. Streaming STT row named Deepgram Flux English, matched to the official API model ID flux-general-en on /v2/listen. Ocular provides eight licensed recordings from four real phone conversations, with eight usable model outputs. Cekura labels the uncertainty estimate a likely 95% range. The endpoint build is not pinned.
- Context
- The published interval is Cekura's likely 95% range, not a separately documented model confidence interval. The page states that per-model counts are usable outputs. This distinct real-call WER is not the existing Coval, AA-WER, or VoiceCodeBench result.
Speko streaming FLEURS test+dev+train English WER, deepgram:flux-general-en (Show measurement, test conditions, and source)
- Source value
- 7.2
- Score
- 7.2
- Metric
- Word error rate on streaming transcript; lower is better
- Unit
- %
- Benchmark version
- Speko STT streaming FLEURS test+dev+train run; 2026-10-01.
- Category
- speech-to-text
- Direction
- Lower is better
- Sample
- 1541
- Source type
- independent-evaluator
- Evaluator
- Speko independent speech benchmark
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Independent evaluation. English read FLEURS clips streamed at real-time pace through the exact Speko provider route; source combines FLEURS test, development and training clips. Speko reports n=1,541 clips and measurement date 2026-10-01.
- Context
- The source explicitly combines FLEURS test, development and training clips, so this is not a clean held-out test set. It does not publish a confidence interval or immutable provider build; its displayed cell text supplies the exact route and evaluated corpus.
Speko EdAcc spontaneous accented English conversation WER, deepgram:flux-general-en (Show measurement, test conditions, and source)
- Source value
- 19.684
- Score
- 19.684
- Metric
- Word error rate over spontaneous accented conversational speech; lower is better
- Unit
- %
- Benchmark version
- Speko EdAcc spontaneous-conversation realtime run; 2026-10-01.
- Category
- speech-to-text
- Direction
- Lower is better
- Sample
- 283
- Source type
- independent-evaluator
- Evaluator
- Speko independent speech benchmark
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Independent evaluation. Spontaneous accented English conversation (EdAcc), 60 speakers, 283 clips, two real-time passes, exact Speko provider route. Speko reports n=283 clips and measurement date 2026-10-01; the two passes are not treated as independent examples.
- Context
- The Speko source displays a range (e.g. 14.4–19.9%) in the same cell but does not label its interval type or confidence level; it is not imported as a confidence interval. It reports 60 speakers and two passes; the N is 283 source clips, not 566 passes. No immutable backend build is published.
WER, report headline (Show measurement, test conditions, and source)
- Source value
- 13.5
- Score
- 13.5
- Metric
- Word error rate (WER)
- Unit
- %
- Benchmark version
- AssemblyAI Universal-3.6 Pro Realtime launch report, published 2026-09-29
- Category
- audio
- Direction
- Lower is better
- Source type
- vendor-reported
- Evaluator
- AssemblyAI Universal-3.6 Pro Realtime report
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Vendor-reported result. AssemblyAI's report says the evaluation used 12,460 scripted caller scenarios voiced by real speakers, three noise environments, and 12 speaker groups. Competitors were called through public streaming APIs at provider defaults with the same audio and scoring pipeline. This is a vendor-reported row; its row-specific denominator is not published. The report labels the evaluated route Flux General English.
- Context
- Vendor-reported, report-specific result. It is not an independently run cross-model cohort; exact API snapshot, per-row N, and confidence intervals are not disclosed unless stated separately.
Entity error rate, report headline (Show measurement, test conditions, and source)
- Source value
- 30.1
- Score
- 30.1
- Metric
- Entity error rate
- Unit
- %
- Benchmark version
- AssemblyAI Universal-3.6 Pro Realtime launch report, published 2026-09-29
- Category
- audio
- Direction
- Lower is better
- Source type
- vendor-reported
- Evaluator
- AssemblyAI Universal-3.6 Pro Realtime report
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Vendor-reported result. AssemblyAI's report says the evaluation used 12,460 scripted caller scenarios voiced by real speakers, three noise environments, and 12 speaker groups. Competitors were called through public streaming APIs at provider defaults with the same audio and scoring pipeline. This is a vendor-reported row; its row-specific denominator is not published. The report labels the evaluated route Flux General English.
- Context
- Vendor-reported, report-specific result. It is not an independently run cross-model cohort; exact API snapshot, per-row N, and confidence intervals are not disclosed unless stated separately.
Mean code-switch WER, report headline (Show measurement, test conditions, and source)
- Source value
- 52.86
- Score
- 52.86
- Metric
- Mean word error rate (WER) on the report's code-switch set
- Unit
- %
- Benchmark version
- AssemblyAI Universal-3.6 Pro Realtime launch report, published 2026-09-29
- Category
- audio
- Direction
- Lower is better
- Source type
- vendor-reported
- Evaluator
- AssemblyAI Universal-3.6 Pro Realtime report
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Vendor-reported result. AssemblyAI's report says the evaluation used 12,460 scripted caller scenarios voiced by real speakers, three noise environments, and 12 speaker groups. Competitors were called through public streaming APIs at provider defaults with the same audio and scoring pipeline. This is a vendor-reported row; its row-specific denominator is not published. The report labels the evaluated route Flux General English.
- Context
- Vendor-reported, report-specific result. It is not an independently run cross-model cohort; exact API snapshot, per-row N, and confidence intervals are not disclosed unless stated separately.
English short-form WER (Show measurement, test conditions, and source)
- Source value
- 6.53
- Score
- 6.53
- Metric
- Word error rate (WER) on English short-form utterances
- Unit
- %
- Benchmark version
- AssemblyAI Universal-3.6 Pro Realtime launch report, published 2026-09-29
- Category
- audio
- Direction
- Lower is better
- Source type
- vendor-reported
- Evaluator
- AssemblyAI Universal-3.6 Pro Realtime report
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Vendor-reported result. AssemblyAI's report says the evaluation used 12,460 scripted caller scenarios voiced by real speakers, three noise environments, and 12 speaker groups. Competitors were called through public streaming APIs at provider defaults with the same audio and scoring pipeline. This is a vendor-reported row; its row-specific denominator is not published. The report labels the evaluated route Flux General English.
- Context
- Vendor-reported, report-specific result. It is not an independently run cross-model cohort; exact API snapshot, per-row N, and confidence intervals are not disclosed unless stated separately.
Page 1 of 20
Popularity compared
How often the model is searched for, used and mentioned, compared with every model of its category.
Popularity index
0 to 100, higher is more popular
20 of 83 models
As of November 2026 · Source: gradually.ai popularity index
Over time
Weekly values from October 18, 2026 to November 6, 2026
FluxEleven v3 39GPT-Live-Transcribe 39GPT-4o Mini TTS 36Gemini 3.5 Transcribe Live 35
Method changes in the timeline, a jump at these points is no market event:
- 1November 1, 2026: Value only with search or usage data, 30-day attention
What is gaining popularity right now, across models, vendors, harnesses and AI subscriptions, is collected in All AI trends.
Flux compared
- 1Flux vs. Nova-324 shared tests
- 2Flux vs. GPT-Live-Transcribe4 shared tests
- 3Flux vs. Gemini 3.5 Transcribe Live17 shared tests
- 4Flux vs. GPT-Realtime-Whisper2 shared tests
- 5Flux vs. Scribe v2 Realtime19 shared tests
- 6Flux vs. Voxtral Mini Transcribe Realtime5 shared tests
- 7Flux vs. Cartesia Ink-25 shared tests
- 8Flux vs. Muse Voice Transcribe3 shared tests
- 9Flux vs. Universal-3.6 Pro Realtime1 shared test
Similar models
More models of the same kind as Flux, chosen by model line, developer, size, price and release date.
- Nova-3Deepgram · same developer
- Aura-2Deepgram · same developer
- Gemini 3.5 TranscribeGoogle · popular in this category
- pyannote Community-1pyannote · popular in this category
- Nemotron-3-DiarizationNVIDIA · popular in this category
- Confucius4-R2T2NetEase Youdao · popular in this category
- NVIDIA Nemotron 3.5 TranscribeNVIDIA · popular in this category
- Granite Speech 5.0 470M TurboCTCIBM · popular in this category
All specifications
Architecture
Input and output
Watermarking and provenance
AI watermarking refers to invisible signals embedded in the output. Visible marks and provenance metadata are listed separately. Unknown means there is insufficient evidence for that particular field.
- AI watermarking
- Unknown
- no source
- Watermarking technique
- Not known
- no source
Licence and access
API
Further details
- Turn detection
- Model-integrated end-of-turn detection with configurable turn-taking
- Source: Turn detection
Sources and data date
Show all 33 sources
Benchmark sources at a glance
Which independent evaluators tested this model. This is coverage, not a quality ranking.
| Evaluator | Test series | Composite indices | Recorded active measurements |
|---|---|---|---|
| Vals AI | 1 | 0 | 3 |
Without vendor claims and historical or disputed results; those stay labelled in the single values.