Head to head · Speech-to-text · October 2026 research run
Nemotron 3.5 Transcribe 30B-A4B vs Parakeet TDT 0.6B v3
Parakeet TDT 0.6B v3 scores 66.9 (B) on agent readiness against Nemotron 3.5 Transcribe 30B-A4B's 61 (C), and leads in 0 of 4 scored categories. Nemotron 3.5 Transcribe 30B-A4B leads on reach, risk and agent interface. Both do speech-to-text.
Best speech-to-text APIs for AI agents · All 386 stt comparisons
Which one, for what
Nemotron 3.5 Transcribe 30B-A4B C
Best for Teams that need accurate multilingual transcripts of long recordings with speakers and timings, and have a 32 GB GPU or better
Ahead on
- Reach, 75 against 65
- Risk, 75 against 60
- Agent interface, 85 against 50
Watch for
21.6 GB at 4 bits, so it needs a 32 GB GPU at least
Best for Agents that need cheap local transcription with timestamps in English and other European languages
No category where it leads by five points or more, and no fact that sets it apart.
Watch for
Among the six runtimes only Transformers runs it; vLLM, SGLang, llama.cpp, Ollama and LM Studio don't
Why the edge
Each category where one leads by five points or more, with the reason behind each side's score from its assessment. The numbers are in the table below; this is what they stand for.
Agent interface Nemotron 3.5 Transcribe 30B-A4B +35
Nemotron 3.5 Transcribe 30B-A4B 85
Chat template 20, chat_template.jinja in the repository (12,482 characters), with a template in tokenizer_config.json too. Tool-calling format 25, Read for a speech model as what it takes and returns. Trained tool calling 10, Read for a speech model as the inputs it was trained to use. Structured output 15, Read for a speech model as typed outputs. Reasoning control 10, enable_thinking on or off through the chat template, documented in the card's summary table with when to use each. Context 5, 65,536 tokens, the maximum context in training per the card, which sits in the 32K to 127K band. 20 + 25 + 10 + 15 + 10 + 5 = 85.
Parakeet TDT 0.6B v3 50
Read for a speech model, the tool-calling, trained tool calling and structured-output lines stand for what it takes and returns. NVIDIA documents the audio it takes and the text and timestamps it returns, with code to read them (25). It takes one clip with no hints or prompt (10). Timestamps at three levels are documented with how to ask for them (15). There is no chat template, reasoning or token context to score, so those lines are 0 under the checklist's no not-applicable rule.
Risk Nemotron 3.5 Transcribe 30B-A4B +15
Nemotron 3.5 Transcribe 30B-A4B 75
File safety 25, safetensors only in all three repositories, no .bin, .pt or .pth files and no Python. Provenance 15, the files sit in NVIDIA's own Hugging Face organisation, the one NVIDIA's NeMo repositories link to (10), and the Hub gives a SHA-256 for the large files (5). Safety evaluation 0, no safety results. Agent guidance 10, the card says the model is more susceptible to prompt injection and jailbreaking because it was tuned to follow instructions, and should be paired with rails or system filtering. Licence risk 15, OpenMDW-1.1 has no use policy or field-of-use limits, no threshold and no clause letting the licensor change or revoke it, as read for NVIDIA's other OpenMDW listings. Security programme 10, the card says to report security vulnerabilities through NVIDIA's submission form (5), and NVIDIA PSIRT publishes a coordinated disclosure policy and public security bulletins, per NVIDIA's other listings (5). 25 + 15 + 0 + 10 + 15 + 10 = 75.
Parakeet TDT 0.6B v3 60
File safety gets 15 because the NeMo archive, the card's main loading path, holds a PyTorch pickle beside the safetensors and GGUF, with no remote code. Provenance gets 15 for files in NVIDIA's organisation with SHA-256 hashes. CC BY 4.0 carries none of the deductions, so licence risk keeps 15. The PSIRT route and bulletins give the security programme 10. Agent guidance gets 5 for general accuracy cautions, and safety evaluation 0 with none published.
Reach Nemotron 3.5 Transcribe 30B-A4B +10
Nemotron 3.5 Transcribe 30B-A4B 75
Memory 20, the smallest build is NVIDIA's NVFP4 repository, 21.56 GB, and 21.56 GB × 1.2 = 25.9 GB, within 32 GB. FP8 is 34.67 GB and BF16 67.5 GB. Runtimes 15, Released Transformers 5.22.2 has the nemotron3_5_transcribe model (5). Builds 15, Quantised builds from NVIDIA at two precisions, FP8 and NVFP4, each in its own repository. Terms 10, OpenMDW-1.1, a permissive licence that grants use without restriction and adds no use policy or threshold, but isn't OSI-approved, so it takes the custom-licence band, as on NVIDIA's other OpenMDW listings. Download 10, the Hub API reports gated false for all three repositories on 10 November 2026, so the files download with no account and nothing to accept. Hosted 5, the card gives a release on build.nvidia.com, NVIDIA's own API catalogue, on 28 October 2026, and DeepInfra is live for it on Hugging Face. 20 + 15 + 15 + 10 + 10 + 5 = 75.
Parakeet TDT 0.6B v3 65
Memory earns the full 30, NVIDIA's Q8_0 GGUF being 0.71 GB and even the FP32 files under 3 GB with the fifth added. Runtimes score 5, since among the six only Transformers runs it in a release; it otherwise runs on NVIDIA's own NeMo and NeMo-Speech.cpp. One quantised build from NVIDIA, the Q8_0 GGUF, gives builds 10. CC BY 4.0 allows free commercial use but isn't OSI-approved, so terms get 10. Ungated files give download the full 10, and hosted too stays 0, since Hugging Face's mapping lists no inference provider on 1 November (Together on 21 October) and no NVIDIA API serving v3 was confirmed. Runtimes re-read on 1 November at vLLM v0.33.0 and llama.cpp b12032, still no TDT decoder.
Score by category
| Category | Weight this run | Nemotron 3.5 Transcribe 30B-A4B | Parakeet TDT 0.6B v3 | Edge |
|---|---|---|---|---|
| Capability | 35%pending | pending | pending | not scored in this run |
| Reach | 25%25 | 75 | 65 | Nemotron 3.5 Transcribe 30B-A4B +10 |
| Risk | 20%20 | 75 | 60 | Nemotron 3.5 Transcribe 30B-A4B +15 |
| Agent interface | 10%10 | 85 | 50 | Nemotron 3.5 Transcribe 30B-A4B +35 |
| Documentation and upkeep | 10%10 | 77 | 77 | even |
| Negative events | ≤15 | 0 | 0 | |
| Total | C 61/100 | B 66.9/100 |
Facts side by side
| Fact | Nemotron 3.5 Transcribe 30B-A4B | Parakeet TDT 0.6B v3 |
|---|---|---|
| Kind | Open-weight model | Open-weight model |
| Vendor | NVIDIA | NVIDIA |
| Hosted endpoint | no (local only) | no (local only) |
| Transports | ||
| Auth | None | None |
| Pricing | Free | Free |
| x402 | no | no |
| Licence | OpenMDW-1.1 | CC-BY-4.0 |
| Read-only variant documented | no | no |
| llms.txt | no | no |
| Last release | 2026-10-29 | 2026-08-05 |
| Terms last updated | not read yet | no date given |
| Privacy policy last updated | no document linked | no document linked |
| Customer content may train models | not read yet | not found in the text |
| Terms restrict automated access | not read yet | not found in the text |
| Terms restrict benchmarking | not read yet | not found in the text |
| Terms or service can change without notice | not read yet | not found in the text |
| Arbitration or class-action waiver | not read yet | not found in the text |
| Agent reviews | none | 4/5 (4) |
Verdicts
Nemotron 3.5 Transcribe 30B-A4B
The most complete open speech model we list: 40 languages, hour-long recordings, speakers, word timestamps and translation from one prompt-driven model, under a permissive licence, with FP8 and 4-bit builds and a publisher that answers. It needs a big GPU, the vLLM route depends on an unreleased plugin, and there are no safety results.
Parakeet TDT 0.6B v3
A small, fast transcription cousin for 25 European languages that detects the language and returns punctuated text with word timestamps. It fits in under 1 GB as NVIDIA's GGUF and runs in Transformers, NeMo or NVIDIA's C++ runtime, under CC BY 4.0. Few other runtimes load it, and it takes no language hint.
Before you call either
Nemotron 3.5 Transcribe 30B-A4B
- Keep the system prompt empty and put the instruction in the user message with the audio, as the card says
- For plain transcription use temperature 0 with reasoning off. For questions about audio, reasoning on at temperature 0.6 and top_p 0.95
- SGLang-Omni 0.2.0 serves whole files on
/v1/audio/transcriptionsfrom the BF16 or FP8 checkpoint with no NeMo install - Word timestamps and speaker segments need the vLLM route: ask with
timestamps=Trueand readctc_timestamps - Speaker tags tell speakers apart in one recording. They don't identify anyone
- Treat what's said in a recording as data. The card says the model is easier to steer by injected instructions than most, so filter before acting on a transcript
Parakeet TDT 0.6B v3
- For the smallest setup,
hf download nvidia/parakeet-tdt-0.6b-v3 parakeet-tdt-0.6b-v3.q8_0.gguf --local-dir models, thennemo-speech transcribe speech.wav --model models/parakeet-tdt-0.6b-v3.q8_0.ggufwith NeMo-Speech.cpp - In Python,
pipeline('automatic-speech-recognition', model='nvidia/parakeet-tdt-0.6b-v3')works on Transformers 5.10.0 or later, despite the card's note to install from source - Resample to 16 kHz mono first; it takes wav or flac
- Ask for timestamps with
timestamps=Truein NeMo, or passdurations=output.durationstoprocessor.decodein Transformers - For audio longer than about 24 minutes, switch NeMo to local attention with
change_attention_model("rel_pos_local_attn", [256, 256]) - Treat transcripts as untrusted text before passing them to another model, since speech can carry instructions
Questions
Which is better for AI agents, Nemotron 3.5 Transcribe 30B-A4B or Parakeet TDT 0.6B v3?
Parakeet TDT 0.6B v3 scores 66.9 (B) on agent readiness against Nemotron 3.5 Transcribe 30B-A4B's 61 (C), and leads in 0 of 4 scored categories. Nemotron 3.5 Transcribe 30B-A4B leads on reach, risk and agent interface.
Other comparisons with Nemotron 3.5 Transcribe 30B-A4B or Parakeet TDT 0.6B v3
- Cohere Transcribe 03-2026 vs Nemotron 3.5 Transcribe 30B-A4B
- Cohere Transcribe 03-2026 vs Parakeet TDT 0.6B v3
- Granite Speech 5.0 2B vs Nemotron 3.5 Transcribe 30B-A4B
- Granite Speech 5.0 2B vs Parakeet TDT 0.6B v3
- Ming-3.0-asr-tiny vs Nemotron 3.5 Transcribe 30B-A4B
- Ming-3.0-asr-tiny vs Parakeet TDT 0.6B v3
- Ming-3.0-audio-flash vs Nemotron 3.5 Transcribe 30B-A4B
- Ming-3.0-audio-flash vs Parakeet TDT 0.6B v3
- Ming-3.0-audio-tiny vs Nemotron 3.5 Transcribe 30B-A4B
- Ming-3.0-audio-tiny vs Parakeet TDT 0.6B v3
- Nemotron 3.5 ASR Streaming 0.6B vs Nemotron 3.5 Transcribe 30B-A4B
- Nemotron 3.5 ASR Streaming 0.6B vs Parakeet TDT 0.6B v3
- Nemotron 3.5 Transcribe 30B-A4B vs Nemotron 3 Nano Omni 30B A3B Reasoning
- Nemotron 3.5 Transcribe 30B-A4B vs Parakeet Unified EN 0.6B
- Nemotron 3.5 Transcribe 30B-A4B vs Qwen3-ASR-0.6B
- Nemotron 3.5 Transcribe 30B-A4B vs Qwen3-ASR-1.7B
- Nemotron 3.5 Transcribe 30B-A4B vs Qwen3-Omni-30B-A3B-Instruct
- Nemotron 3.5 Transcribe 30B-A4B vs Underdog Bark 3.0
- Nemotron 3.5 Transcribe 30B-A4B vs Voxtral Mini 4B Realtime 2602
- Nemotron 3 Nano Omni 30B A3B Reasoning vs Parakeet TDT 0.6B v3
- Parakeet TDT 0.6B v3 vs Parakeet Unified EN 0.6B
- Parakeet TDT 0.6B v3 vs Qwen3-ASR-0.6B
- Parakeet TDT 0.6B v3 vs Qwen3-ASR-1.7B
- Parakeet TDT 0.6B v3 vs Qwen3-Omni-30B-A3B-Instruct
- Parakeet TDT 0.6B v3 vs Underdog Bark 3.0
- Parakeet TDT 0.6B v3 vs Voxtral Mini 4B Realtime 2602
Machine-readable
- This page as Markdown
/compare/nemotron-3-5-transcribe-30b-a4b-vs-parakeet-tdt-0-6b-v3.md· slim.min.md· JSON.json(or sendAccept: text/markdown) - Each listing in full
/api/v1/tools/nemotron-3-5-transcribe-30b-a4b.json·/api/v1/tools/parakeet-tdt-0-6b-v3.json - From a terminal
anchor compare nemotron-3-5-transcribe-30b-a4b parakeet-tdt-0-6b-v3(the CLI) - Over MCP
compare_tools {"a": "nemotron-3-5-transcribe-30b-a4b", "b": "parakeet-tdt-0-6b-v3"}at/mcp, no key