Data as of November 12, 2026 · How the score is built
Released Nov 12, 2026N/A contextSalesforce model card
SOV-Transcribe-1
Decision readingSOV-Transcribe-1 is tracked, but not publicly ranked yet. Published speech or voice results from SOV-Transcribe-1 launch evaluation appear below. They do not produce a weighted text-model score.
Released Nov 12, 2026 — see all recent releases
Email me when its price, rank or availability changes.
Voice benchmark evidence
SOV-Transcribe-1 launch evaluation is an owner-defined voice evaluation. Its result stays separate from BenchLM's weighted text-model ranking.
Independent test: Independent speech-to-text board
Independent test by an outside speech-to-text board: batch transcription of English voice-agent, parliament and earnings-call audio through each provider's API, read 2026-11-12. The index weights the three sets 50/25/25; error rates are lower-is-better.
Lineage
The sequence follows explicit supersedes links. Each score is estimated for that model; a relative can inform a sparse estimate but never sets a floor, so a newer release can score below an earlier one. Scores and prices remain blank when the corresponding public row or first-party rate is unavailable.
Base entry
Spec sheet
Each documented value carries its source. Missing fields stay visible as not sourced or not published, rather than disappearing from the page.
- API model ID
- salesforce-sovereign-intelligence/SOV-Transcribe-1Salesforce model documentation
- Context window
- N/A
- Maximum output
- Not sourced yet
- Knowledge cutoff
- Not sourced yet
- Input modalities
- Not sourced yet
- Output modalities
- Not sourced yet
- Parameters
- Not sourced yet
- Availability
- Weights published on Hugging Face under Apache 2.0 (model card dated November 12, 2026), in BF16 with the MoE router in FP32. It takes 16 kHz audio up to 30 minutes (tested) and runs with Transformers (remote code) or a vLLM 0.32.0 plugin shipped in the repository, which serves an OpenAI-compatible transcription endpoint; native vLLM and Transformers support is being upstreamed. An optional DFlash draft model in its own repository speeds up decoding about 2–3× for one request at a time. It is batch transcription, not a streaming service.
- Cloud regions
- Not tracked yet
- Lifecycle
- Current
- API capabilities
- Tool calling, structured outputs, and batch support are not tracked yet
- Prompt caching
- Not documented in the pricing recordSalesforce model documentation
- Self-host
- Open weights available; hardware estimate not sourced
- Rate limits
- Not tracked yet
How to read this profile
The visual layer above carries the decisions. These notes preserve the model, ranking, coverage, and family context behind the numbers.
We track SOV-Transcribe-1, but no weighted text-model benchmark result is published on the site yet. This page shows the metadata and separate protocol evidence we can verify now; a BenchLM score will appear only if compatible public evaluations land.
SOV-Transcribe-1 is a open weight model with a N/A context window. No explicit reasoning mode is documented in this profile.
Weights published on Hugging Face under Apache 2.0 (model card dated November 12, 2026), in BF16 with the MoE router in FP32. It takes 16 kHz audio up to 30 minutes (tested) and runs with Transformers (remote code) or a vLLM 0.32.0 plugin shipped in the repository, which serves an OpenAI-compatible transcription endpoint; native vLLM and Transformers support is being upstreamed. An optional DFlash draft model in its own repository speeds up decoding about 2–3× for one request at a time. It is batch transcription, not a streaming service.
SOV-Transcribe-1 is Salesforce’s open-weight speech-augmented language model for non-streaming speech recognition: a Whisper large-v3 audio encoder and projector feeding Gemma 4 26B A4B, a mixture-of-experts decoder, for 26.6B parameters with 4.6B active. It transcribes English, French, German, Spanish, Italian, Portuguese and Dutch, verbatim or cleaned up, and takes plain-language instructions for style plus names, terms or a topic as context. Its provider-reported speech results stay separate from text-model rankings.
Speech or voice evidence appears above; text-model benchmark coverage remains empty.
Last updated November 12, 2026. Runtime fields remain blank until a sourced snapshot exists.
Deployment options
Self-host and provider-specific paths stay separate from benchmark evidence so operating constraints are visible before a score becomes the whole decision.
Published weights are available, but BenchLM does not yet have a sourced parameter and VRAM profile for this exact model. Hardware cost estimates stay unavailable until that sizing record is complete.
Estimate VRAM from known parametersQuestions
How does SOV-Transcribe-1 perform overall in AI benchmarks?
SOV-Transcribe-1 has published speech or voice results in SOV-Transcribe-1 launch evaluation. Those results remain separate from the weighted text-model ranking, so this profile does not assign a BenchLM score or overall rank. Use the linked evaluation's metric and run conditions to compare speech systems.
Is SOV-Transcribe-1 open source?
SOV-Transcribe-1 is an open-weight model from Salesforce. Its weights can be downloaded for local or hosted deployment, subject to the published license. Open weight does not automatically mean open source: training data and training code may remain private, and commercial restrictions can still apply.
Does SOV-Transcribe-1 have full benchmark coverage on BenchLM?
No. SOV-Transcribe-1's speech or voice results appear in SOV-Transcribe-1 launch evaluation, but it has no published text-model benchmark rows. The voice evaluation does not fill the text-model coverage slots or produce an overall rank. Missing text coverage is not treated as a zero score.
What is the context window size of SOV-Transcribe-1?
SOV-Transcribe-1 has a reported context window of N/A in the exact-model catalog record. The value stays visible, but the profile marks its source link as unavailable instead of presenting it as directly documented. Maximum output length remains separate because providers often publish a different limit.