selfhost.directory

The index of self-hostable projects — ranked, categorized and served with live updates.
Own your data and stop giving it to big tech.

ServiceNow-AI logo

Apriel-Voice-27B-preview

ServiceNow-AI·Instruct·released 13 Oct 2026

MIT27.2BMoE256K ctx ≈19 GB
Your GPU
hf download ServiceNow-AI/Apriel-Voice-27B-preview
Setup ↓

An open model for speech, reasoning, and tool use from ServiceNow

Notify me when

How much VRAM does Apriel-Voice-27B-preview need?

At Q4_K_M Apriel-Voice-27B-preview needs about 19 GB with an 8K context (19 GB at Q4, 56 GB at full precision). Pick your GPU to see what fits and how fast it runs.

--vram
any
--gpus
--quant

Needs: ≈19 GB · Q4_K_M · 8K context — Pick a GPU to check the fit.

Change quant, context or KV cache to compare
Advanced: context 8K · KV cache 16-bit
Context
KV cache

Smallest GPU for each quant and context (KV cache 16-bit)

Quantisation8K16K32K64K128K256K
Q3_K_M
Q4_K_M
Q5_K_M
Q6_K
Q8_0
FP16

≤16 GB card24–32 GB48–96 GBmore

QuantisationQualityWeightsKV cacheTotal
Q2_KHeavy lossSmallest, noticeably worse qualityHeavy loss8.3 GB 1.9 GB≈11 GB
Q3_K_MNoticeable lossSmall, some quality lossNoticeable loss12 GB 1.9 GB≈15 GB
Q4_K_MrecommendedSmall lossRecommended default for local useSmall loss15 GB 1.9 GB≈19 GB
Q5_K_MNear-losslessNear-lossless for most tasksNear-lossless18 GB 1.9 GB≈21 GB
Q6_KNear-losslessPractically losslessNear-lossless21 GB 1.9 GB≈24 GB
Q8_0LosslessLossless in practice, twice the size of Q4Lossless27 GB 1.9 GB≈31 GB
FP16 / BF16LosslessFull precision, as releasedLossless51 GB 1.9 GB≈56 GB

Estimates for 8K tokens of context: weights + KV cache + runtime overhead. All experts stay in memory, but only the active ones are read per token. Speed assumes memory-bound decoding at your GPU’s bandwidth. Pick a GPU above and the fit column checks every row against it.

How to pick a quant

Take the biggest quant that fits with room left for context. Q4_K_M is the usual sweet spot. Below Q4 quality drops fast, above Q6 you mostly pay memory for nothing. Big models (70B+, large MoE) tolerate low quants better than small ones.

  • LosslessF16, Q8_0
  • Near-losslessQ6_K, Q5_K_M
  • Small lossQ5_K_S, Q4_K_M, IQ4_NL, IQ4_XS, Q4_K_S
  • Noticeable lossQ4_0, Q3_K_L, Q3_K_M, IQ3_M
  • Heavy lossIQ3_XXS, Q3_K_S, IQ2_M, Q2_K

Typical quality loss, not measured per model. Tiers come from KL divergence against FP16 on Llama 3 8B: llama.cpp perplexity README: LLaMA 3 8B scoreboard (KLD and ΔPPL vs FP16, Wikitext-2); llama.cpp quantize tool: ΔPPL per quant type @ Llama-3-8B

How much context fits?

Model quant
KV cache

At 8K context, Q4_K_M needs ≈19 GB(15 GB weights + 1.9 GB KV cache + 1.3 GB overhead). Pick a GPU above to see how much context fits.

0204060801001K2K4K8K16K32K64K128K256KGB

Weights + overheadKV cache

Can the Tesla P40 run Apriel-Voice-27B-preview? →

Speed vs quality per quant

Smaller quants run faster and fit in less memory, but lose quality.

Q2_KQ3_K_MQ4_K_MQ5_K_MQ6_KQ8_0F16

Fits in VRAMTight fitCPU offload✕ won't runQuality is typical KL divergence vs FP16, not measured for this model.

Real-world speed of Apriel-Voice-27B-preview

Tokens per second measured by people running it at home. A GPU shows a median once 3 reports agree on the same setup. Reports far from our estimate wait for a review. Measure yours with llama-bench →

No medians yet. Be one of the first 3 to report a setup.

All reports

Run Apriel-Voice-27B-preview locally

Quantised versions and downloads

RepositoryFormatQuantSizeDownloads
makcimbx/Apriel-Voice-27B-preview-NVFP4SafetensorsNVFP4—66

About Apriel-Voice-27B-preview

An open model for speech, reasoning, and tool use from ServiceNow Apriel Voice 27B Preview understands spoken requests, reasons over conversational context, and generates text responses and tool calls through a shared Gemma 4 backbone. It adds a speech interface while retaining text and image input capabilities. Pair it with an external speech synthesizer to build interactive voice assistants, customer support workflows, and spoken tool use. Try Apriel Live · Read the blog · Technical report The model supports transcription, speech translation, spoken instruction following, reasoning, and function calling. The application executes tools and returns their results to the conversation. Real-time voice interaction is supported by the application and serving stack; the checkpoint itself generates text and tool calls rather than native full-duplex speech. A Whisper large-v3-turbo speech encoder connects to the Gemma 4 26B-A4B backbone through a learned projector. Training first aligns this speech interface, then adapts it with the backbone using speech tasks and text replay. Cross-entropy and teacher-distillation supervision vary by data source; the technical report describes the full recipe.

Full model card on Hugging Face →

Alternatives to Apriel-Voice-27B-preview

All alternatives →

NVIDIA Nemotron 3 Nano Omni is a multimodal large language model that unifies video, audio, image, and text understanding to support enterprise-grade Q&A, summarization, transcription, and document intelligence workflows.

33BA3B256K ctx nvidia-open-model-agreement
93.9KApr 2026gguf
Qwen logo≈22 GB

Qwen3-Omni is the natively end-to-end multilingual omni-modal foundation models. It processes text, images, audio, and video, and delivers real-time streaming responses in both text and natural speech. We introduce several architectural upgrades to improve performance and efficiency. Key features:

35.3BA3B64K ctx Apache 2.0
340.5KSept 2025gguf
Qwen logo≈20 GB

Qwen3-Omni is the natively end-to-end multilingual omni-modal foundation models. It processes text, images, audio, and video, and delivers real-time streaming responses in both text and natural speech. We introduce several architectural upgrades to improve performance and efficiency. Key features:

31.7BA3B64K ctx Apache 2.0
150KSept 2025gguf
google logo

gemma-4-12B-it

google·Omni

≈11 GB

Gemma is a family of open models built by Google DeepMind. Gemma 4 models are multimodal, handling text and image input (with audio supported on E2B, E4B, and 12B) and generating text output. This release includes open-weights models in both pre-trained and instruction-tuned variants.

12B256K ctx Apache 2.0
1.6MMay 2026ollama

NVIDIA-NemotronLabs-AI-for-Media-Sports-Tennis is a fine-tuned multimodal large language model derived from Nemotron 3 Nano Omni. It was further trained on an in-house NVIDIA tennis dataset of manually labeled point-level video clips to specialize in tennis match understanding.

33B256K ctx openmdw-1.1
2.4KSept 2026gguf

NVIDIA-NemotronLabs-AI-for-Media-Sports-Basketball is a fine-tuned multimodal large language model derived from Nemotron 3 Nano Omni.

33B256K ctx openmdw-1.1
1.2KSept 2026

Apriel-Voice-27B-preview FAQ

How much VRAM does Apriel-Voice-27B-preview need?
About 19 GB at Q4_K_M with an 8K context. Q8_0 needs about 31 GB and full precision about 56 GB. Longer contexts need more memory for the KV cache.
Can Apriel-Voice-27B-preview run on the Tesla P40?
Yes. It fits at Q5_K_M (about 21 GB), at an estimated 9.9 tokens/s. It fits comfortably on the Radeon RX 7900 XTX, Instinct MI50 32GB, RTX 5090.
Can I use Apriel-Voice-27B-preview commercially?
Yes. It is released under MIT. Permissive: do almost anything as long as you keep the copyright notice.
What is the context length of Apriel-Voice-27B-preview?
262,144 tokens (256K). The full window needs far more memory than an 8K context.

Data from Hugging Face, the model card and public leaderboards. Memory and speed figures are estimates; real numbers depend on the runtime, context length and drivers. Always check the licence yourself before commercial use.