Gemini 3.5 vs Soniox v5 vs Linden-1: STT Guide 2026

Google made Gemini 3.5 Transcribe and Transcribe Live generally available on August 26, 2026, joining Soniox's stt-rt-v5 (released June 16) and Speechmatics Linden-1, which LiveKit Agents added on August 27, in competing for production voice-agent speech-to-text workloads. A comparison published August 29, 2026 notes the three models differ sharply on session limits — Gemini live sessions cap at 10 minutes, Soniox at 300 minutes, and Linden-1 has no published Linden-specific limit — and on pricing, with public reference estimates of about $0.009/min for Gemini direct, $0.0095/min via LiveKit, $0.002/min for Soniox, and $0.005/min for Linden-1. The comparison states no neutral, reproducible test currently ranks these three exact models on the same German 8 kHz telephone dataset, so it recommends a controlled A/B test on the user's own calls rather than treating published prices as a quality ranking.

Summarize Content With: Three new or recently updated speech-to-text options now compete for production voice-agent workloads. Google made Gemini 3.5 Transcribe and Transcribe Live generally available on August 26, 2026, Soniox released stt-rt-v5 on June 16, and LiveKit Agents added Gemini 3.5 Transcribe Live and Speechmatics Linden-1 on August 27 Google changelog https://ai.google.dev/gemini-api/docs/changelog , Soniox model documentation https://soniox.com/docs/stt/models , LiveKit Agents releases https://github.com/livekit/agents/releases . This is a research-based comparison, not a lab benchmark. No neutral, reproducible test currently ranks these three exact models on the same German 8 kHz telephone dataset. The useful question is therefore not "Which model wins overall?" It is which access path, session limit, telephone format, endpointing behavior, language handling, data route, and billing model fit your calls. If you need the basics first, start with what speech-to-text means https://www.famulor.io/glossary/stt-speech-to-text . Key Takeaways - Gemini offers broad direct multilingual coverage and native LiveKit Inference access, but live sessions are limited to ten minutes and the LiveKit route cannot be regionally pinned. - Soniox combines long sessions, telephone-native codecs, live diarization, code-switching, and configurable semantic endpointing through a plugin that uses your own Soniox key. - Linden-1 is a native LiveKit Inference option with server-side end-of-speech detection, but its local vad input is ignored and no Linden-specific regional guarantee is published.- Public prices describe different access paths and billing units. They are not a quality ranking. - A decision should follow a controlled A/B test on your own calls, especially German public telephone network PSTN audio, entity capture, turn boundaries, and actual cost per conversation minute. Quick comparison | Criterion | Gemini 3.5 Transcribe Live | Soniox stt-rt-v5 | Speechmatics Linden-1 | |---|---|---|---| | Exact integration ID | google/gemini-3.5-transcribe-live in LiveKit | model="stt-rt-v5" in the Soniox plugin | speechmatics/linden-1 in LiveKit | | Access path | Google Live API or native LiveKit Inference | Python or Node.js plugin with a Soniox key | Native LiveKit Inference | | Published language scope | Google direct: 85+; LiveKit page: 26 codes | 60+ | 55+ | | Automatic language switching | Yes, including within a live session | Yes, including within a sentence | Multilingual support documented; arbitrary automatic code-switching is not clearly documented | | Live diarization | No | Yes | Available in Speechmatics configuration; verify the Linden path in your setup | | Turn boundary | Google server voice activity detection VAD , with direct hybrid and manual strategies | Semantic endpointing with configurable controls | Server-side end-of-speech; local vad is ignored | | Maximum published session | 10 minutes | 300 minutes | No Linden-specific limit published in the reviewed material | | Telephone input | Direct live API specifies 16-bit PCM, 16 kHz, mono | PCM, mu-law, A-law, and other formats | Media handling is abstracted by LiveKit; no Linden-specific raw codec list published | | Data-region caveat | Global LiveKit deployment, not regionally pinnable | EU endpoints available after residency access is enabled | No Linden-specific LiveKit region guarantee published | | Public reference price | Direct estimate about $0.009/min; LiveKit $0.0095/min | Token-based estimate about $0.002/min | LiveKit $0.005/min | The table is a snapshot dated August 29, 2026. It deliberately avoids accuracy scores because vendor benchmarks do not use a shared dataset, telephone channel, language mix, or endpointing setup. Access and integration paths The three names hide different deployment models that affect credentials, routing, and invoicing. A broader checklist for choosing an STT provider https://www.famulor.io/blog/how-to-choose-the-right-speech-to-text-stt-provider-for-your-voice-ai-agent helps separate provider-level requirements from model-level tests. Gemini 3.5 Transcribe Live The direct model ID is gemini-3.5-transcribe-live ; LiveKit exposes it as google/gemini-3.5-transcribe-live . It supports interim and final results over a live WebSocket session. Google's direct documentation also describes strict transcription and Smart Transcription behavior, plus automatic, hybrid, and manual voice-activity strategies Gemini model page https://ai.google.dev/gemini-api/docs/models/gemini-3.5-transcribe , live transcription guide https://ai.google.dev/gemini-api/docs/live-api/live-transcribe . The important operational constraint is the ten-minute session maximum. Long calls need a planned rollover, overlap, and transcript reconciliation strategy. Gemini live transcription also lacks word timestamps and live speaker diarization. The LiveKit Gemini page https://docs.livekit.io/agents/models/stt/gemini/ currently exposes a narrower language-code list than Google's direct documentation and does not establish that every direct VAD control is configurable through LiveKit Inference. Soniox stt-rt-v5 Soniox uses the LiveKit plugin path rather than LiveKit Inference, so the application supplies its own Soniox key. Set stt-rt-v5 explicitly. Current Node.js plugin documentation may still show v4 as a default, while Soniox aliases v4 to v5 server-side. The Soniox pipeline update https://www.famulor.io/blog/famulor-takes-off-revolutionary-upgrades-for-your-ai-telephony-with-soniox-gpt-realtime-15-and-the-latest-openai-models provides related context. Soniox documents real-time streaming sessions of up to 300 minutes real-time transcription https://soniox.com/docs/stt/rt/real-time-transcription , limits and quotas https://soniox.com/docs/stt/rt/limits-and-quotas . This is useful for long consultations, but teams still need reconnect and failure handling. Speechmatics Linden-1 Linden-1 is accessed as speechmatics/linden-1 through LiveKit Inference, which handles model routing and usage billing. A separate public direct-API SKU for this exact model was not verified. Speechmatics positions Linden for voice agents, and LiveKit recommends it for most Speechmatics agent workloads Speechmatics announcement https://www.speechmatics.com/company/articles-and-news/speechmatics-linden-livekit-inference , LiveKit Speechmatics page https://docs.livekit.io/agents/models/stt/speechmatics/ . End-of-speech is detected server-side. A local LiveKit vad object is ignored for Linden, so teams should not assume that swapping local VAD libraries will change its turn boundary. German and multilingual conversations All three options include German, but their published language claims are not directly equivalent. Google states 85+ languages for the direct model, while LiveKit currently lists 26 language codes for its Gemini integration. Gemini can change languages during a session. Soniox documents 60+ languages and automatic code-switching, including switches within a sentence Google live guide https://ai.google.dev/gemini-api/docs/live-api/live-transcribe , Soniox supported languages https://soniox.com/docs/stt/concepts/supported-languages . Linden's LiveKit page lists 55+ languages, but the reviewed sources do not clearly confirm arbitrary automatic code-switching. For German production traffic, test names, compound nouns, dialects, accented German, and German-English switches. Add domain terms with Gemini custom vocabulary, Soniox context and terms, or Speechmatics additional vocab . Test vocabulary against real utterances using these guidelines for specialist terms and names https://www.famulor.io/blog/custom-vocabulary-for-ai-voice-agents-pronunciation-guide . Turn detection matters beyond transcript accuracy Voice agents act when the STT system decides a user has finished. A correct transcript that arrives after a long pause can still create a poor conversation, while an early endpoint can cut off a date, address, or correction. A current Hacker News discussion https://news.ycombinator.com/item?id=49468818 is useful only as a practitioner signal that turn-taking deserves production testing. Its comments and reported latency figures are not a controlled comparison of these models. Google documents automatic server VAD, a hybrid strategy that combines server speech-start detection with the local audio stream end signal, and fully manual activity markers. The direct hybrid mode can avoid waiting for the server timeout, but configuration parity through LiveKit should be checked rather than assumed. Soniox explicitly describes semantic endpointing based on pauses, intonation, speech patterns, and context. It exposes endpoint latency adjustment level from 0 to 3, endpoint sensitivity from -1.0 to 1.0, and max endpoint delay ms from 500 to 3,000 ms. More aggressive settings can increase segmentation and slightly affect word and diarization accuracy, according to the Soniox endpoint detection documentation https://soniox.com/docs/stt/rt/endpoint-detection . Linden detects the speech end on the server. Do not call this semantic endpointing without a Linden-specific primary source. The architectural consequences are easier to assess alongside the pipeline and realtime architecture guide https://www.famulor.io/blog/realtime-vs-pipeline-voice-agent-architecture-guide-2026 and the voice-agent latency breakdown https://www.famulor.io/blog/ai-voice-agent-latency-how-fast-your-phone-bot-must-reply . Telephone audio and structured data Soniox documents direct carrier-audio support because its real-time API accepts PCM, mu-law, A-law, and other formats. Gemini's direct live-transcription specification calls for 16-bit PCM at 16 kHz mono, which can require resampling from an 8 kHz PSTN stream. LiveKit abstracts media transport for Linden, but that does not prove a Linden-specific raw-codec capability. Score structured fields separately: names, spelled email addresses, telephone numbers, IBANs, appointment /use-cases/appointment-booking-faqs times, postal codes, and product codes. Measure entity error rate as well as word error rate. Gemini has no live diarization, Soniox does, and Speechmatics diarization should be validated on the exact Linden route. Privacy and processing regions No single EU-hosting statement covers all three paths. LiveKit Inference https://docs.livekit.io/agents/models/inference/ is zero data retention by default, but Gemini 3.5 Transcribe Live runs through one global deployment and cannot be selected for a region-pinned request. That prevents a strict regional-only claim for this route. Soniox documents dedicated EU domains, including stt-rt.eu.soniox.com , regional processing, and regional storage. EU data residency must be enabled first Soniox data residency https://soniox.com/docs/data-residency . No Linden-specific region guarantee through LiveKit was verified. Confirm the route contractually before an EU-only rollout and review other voice AI STT deployment choices https://www.famulor.io/blog/8-alternatives-to-azure-for-voice-ai-stt . Price comparison by access route | Access route | Public reference | Important caveat | |---|---|---| | Soniox direct plugin | About $0.002/min, or $0.12/hour | Token-based estimate using Soniox reference consumption; speech share, output, and added context can change cost | | Linden-1 via LiveKit Inference | $0.005/min | Excludes the LiveKit plan, SIP and media, LLM, TTS, storage, taxes, and implementation | | Gemini direct | About $0.009/min blended | Estimate combining $0.005 audio input with roughly $0.004 text output per minute, not a fixed all-in minute rate | | Gemini via LiveKit Inference | $0.0095/min | Excludes the same surrounding voice stack and operational costs | The Soniox estimate follows its token pricing and reference consumption https://soniox.com/pricing . Gemini's estimate follows Google's Gemini API pricing https://ai.google.dev/gemini-api/docs/pricing . The two LiveKit figures come from LiveKit Inference pricing https://livekit.com/pricing/inference . Compare invoices only after normalizing what each route includes. Lower STT cost does not imply lower total cost or higher recognition quality. A neutral A/B test for real calls Use identical audio, downstream logic, and measurement windows. Include: 1. German 8 kHz mu-law PSTN recordings, plus the resampled version required by an integration. 2. Standard German, regional dialects, non-native speakers, and German-English code-switching. 3. Names, email addresses, IBANs, telephone numbers, times, dates, and product codes. 4. Silence, thinking pauses, self-corrections, overlapping speech, background noise, and barge-in. 5. Short service calls and calls longer than ten minutes, including a controlled Gemini session rollover. Report word and entity error rates, time from the final speech frame to a stable transcript, early and late endpoints, partial churn, barge-in errors, reconnects, and actual cost per conversation minute. Review by scenario because one average can hide failures on business-critical fields. Decision matrix | Your main requirement | First option to test | Why it belongs in the test | |---|---|---| | Broad direct multilingual coverage and native LiveKit Inference | Gemini 3.5 Transcribe Live | Direct documentation covers 85+ languages, with 26 codes currently listed by LiveKit | | Long calls and telephone-native codecs | Soniox stt-rt-v5 | 300-minute published limit plus mu-law and A-law support | | Tunable semantic endpointing | Soniox stt-rt-v5 | Explicit semantic controls with documented tradeoffs | | LiveKit-managed access and billing | Linden-1 or Gemini | Both are native LiveKit Inference routes | | Live diarization | Soniox first; Linden route second | Soniox support is explicit; validate the Speechmatics configuration used | | Strict region pinning | Verify before selection | Gemini LiveKit does not qualify; Soniox needs EU residency enablement; Linden needs route confirmation | This matrix prioritizes a test order, not a universal winner. Also verify whether each exact model is available in your current workspace before rollout. Frequently asked questions Which model is most accurate for German telephone calls? There is no neutral head-to-head benchmark for these exact versions on the same German 8 kHz telephone dataset. Test your own calls and report both word error rate and entity error rate. Vendor scores from different datasets cannot establish a fair ranking. Can Gemini 3.5 Transcribe Live handle calls longer than ten minutes? Not as one uninterrupted documented session. The published maximum is ten minutes, so longer calls require a rollover design with overlap, transcript reconciliation, and failure testing. Is Soniox stt-rt-v5 automatically processed in the EU? No. Soniox offers dedicated EU endpoints and regional storage, but data-residency access must be enabled. Confirm the endpoint, storage behavior, and contract before production. Does Linden-1 use semantic endpointing? The reviewed primary sources confirm server-side end-of-speech detection and that local LiveKit VAD is ignored. They do not confirm semantic endpointing for Linden-1, so that label should not be used without newer model-specific documentation. Conclusion and next step Gemini 3.5 Transcribe Live, Soniox stt-rt-v5 , and Speechmatics Linden-1 solve different deployment problems. Gemini emphasizes direct multilingual reach and native LiveKit access, Soniox exposes telephone and endpointing controls through a provider plugin, and Linden offers a LiveKit-managed route with server-side turn completion. None of those differences proves an overall accuracy winner. Choose a shortlist from the decision matrix, verify model availability in your workspace, and run the same German call set through every candidate. If you want to evaluate that configuration against real conversations, book a personal Famulor demo https://www.famulor.io/demo to review your call flow, languages, and telephony environment together. About the author Sarah Müller is the named author of this Famulor comparison. The analysis uses public primary documentation and separates published facts from estimates and practitioner signals. For editorial or product questions, contact Famulor Support https://www.famulor.io/support . Famulor's handling of personal data is described in its privacy information https://www.famulor.io/privacy . Writer at Famulor