Voxtral-Small-24B-2507: transcribe.cpp GGUF

GGUF conversions of mistralai/Voxtral-Small-24B-2507 for use with transcribe.cpp.

Ported from upstream commit da5b424, pinned 2026-06-05. Validated against the Transformers reference at transcribe.cpp commit dac22fa on 2026-06-05.

Offline audio-LLM speech-to-text and speech translation. A Whisper-large-v3 bidirectional audio encoder feeds a 4-frame-group projector (375 audio tokens per 30 s chunk) into a Mistral-Small-24B causal LM (40 layers, GQA 32/8, NEOX RoPE, SwiGLU) via audio-token injection. Takes a 16 kHz mono WAV and produces a transcript via greedy decoding. The larger sibling of Voxtral Mini 3B — same encoder, projector, frontend, and tokenizer, with a scaled-up decoder.

Downloads

WER measured on the full LibriSpeech test-clean split (2620 utterances) with the Whisper English text normalizer, greedy decoding, batch size 8 on an NVIDIA A100 80 GB. Same-machine HuggingFace transformers reference (VoxtralForConditionalGeneration, BF16, greedy): 1.57%; the BF16 GGUF matches at 1.56%. Validation for this variant is end-to-end by WER — the family's tensor-level numerical parity is established by the Voxtral Mini 3B sibling (identical architecture).

Usage

Build transcribe.cpp from source:

git clone git@github.com:handy-computer/transcribe.cpp.git
cd transcribe.cpp
cmake -B build && cmake --build build

Run on a 16 kHz mono WAV:

build/bin/transcribe-cli \
  -m Voxtral-Small-24B-2507-Q8_0.gguf \
  input.wav

If your audio isn't already 16 kHz mono WAV, convert it first:

ffmpeg -i input.mp3 -ar 16000 -ac 1 output.wav

See the transcribe.cpp model page for performance numbers, numerical validation, and reproduction steps.

License

Inherited from the base model: Apache-2.0. See the upstream model card for full terms.

---

Original Model Card

The section below is reproduced from mistralai/Voxtral-Small-24B-2507 at commit da5b424 for offline reference. The upstream card is the authoritative source.

Voxtral Small 1.0 (24B) - 2507

Voxtral Small is an enhancement of Mistral Small 3, incorporating state-of-the-art audio input capabilities while retaining best-in-class text performance. It excels at speech transcription, translation and audio understanding.

Learn more about Voxtral in our blog post here and our research paper.

Key Features

Voxtral builds upon Mistral Small 3 with powerful audio understanding capabilities.

• Dedicated transcription mode: Voxtral can operate in a pure speech transcription mode to maximize performance. By default, Voxtral automatically predicts the source audio language and transcribes the text accordingly
• Long-form context: With a 32k token context length, Voxtral handles audios up to 30 minutes for transcription, or 40 minutes for understanding
• Built-in Q&A and summarization: Supports asking questions directly through audio. Analyze audio and generate structured summaries without the need for separate ASR and language models
• Natively multilingual: Automatic language detection and state-of-the-art performance in the world’s most widely used languages (English, Spanish, French, Portuguese, Hindi, German, Dutch, Italian)
• Function-calling straight from voice: Enables direct triggering of backend functions, workflows, or API calls based on spoken user intents
• Highly capable at text: Retains the text understanding capabilities of its language model backbone, Mistral Small 3.1

Benchmark Results

Audio

Average word error rate (WER) over the FLEURS, Mozilla Common Voice and Multilingual LibriSpeech benchmarks:

Text

Usage

The model can be used with the following frameworks;

• vllm (recommended): See here
• Transformers 🤗: See here

Notes:

• temperature=0.2 and top_p=0.95 for chat completion (e.g. Audio Understanding) and temperature=0.0 for transcription
• Multiple audios per message and multiple user turns with audio are supported
• Function calling is supported
• System prompts are not yet supported

vLLM (recommended)

We recommend using this model with vLLM.

Installation

Make sure to install vllm >= 0.10.0, we recommend using uv

uv pip install -U "vllm[audio]" --system

Doing so should automatically install mistral_common >= 1.8.1.

To check:

python -c "import mistral_common; print(mistral_common.__version__)"

Offline

You can test that your vLLM setup works as expected by cloning the vLLM repo:

git clone https://github.com/vllm-project/vllm && cd vllm

and then running:

python examples/offline_inference/audio_language.py --num-audios 2 --model-type voxtral

Serve

We recommend that you use Voxtral-Small-24B-2507 in a server/client setting.

1. Spin up a server:

vllm serve mistralai/Voxtral-Small-24B-2507 --tokenizer_mode mistral --config_format mistral --load_format mistral --tensor-parallel-size 2 --tool-call-parser mistral --enable-auto-tool-choice

Note: Running Voxtral-Small-24B-2507 on GPU requires ~55 GB of GPU RAM in bf16 or fp16.

2. To ping the client you can use a simple Python snippet. See the following examples.

Audio Instruct

Leverage the audio capabilities of Voxtral-Small-24B-2507 to chat.

Make sure that your client has mistral-common with audio installed:

pip install --upgrade mistral_common\[audio\]

Python snippet

from mistral_common.protocol.instruct.messages import TextChunk, AudioChunk, UserMessage, AssistantMessage, RawAudio
from mistral_common.audio import Audio
from huggingface_hub import hf_hub_download

from openai import OpenAI

# Modify OpenAI's API key and API base to use vLLM's API server.
openai_api_key = "EMPTY"
openai_api_base = "http://<your-server-host>:8000/v1"

client = OpenAI(
    api_key=openai_api_key,
    base_url=openai_api_base,
)

models = client.models.list()
model = models.data[0].id

obama_file = hf_hub_download("patrickvonplaten/audio_samples", "obama.mp3", repo_type="dataset")
bcn_file = hf_hub_download("patrickvonplaten/audio_samples", "bcn_weather.mp3", repo_type="dataset")

def file_to_chunk(file: str) -> AudioChunk:
    audio = Audio.from_file(file, strict=False)
    return AudioChunk.from_audio(audio)

text_chunk = TextChunk(text="Which speaker is more inspiring? Why? How are they different from each other? Answer in French.")
user_msg = UserMessage(content=[file_to_chunk(obama_file), file_to_chunk(bcn_file), text_chunk]).to_openai()

print(30 * "=" + "USER 1" + 30 * "=")
print(text_chunk.text)
print("\n\n")

response = client.chat.completions.create(
    model=model,
    messages=[user_msg],
    temperature=0.2,
    top_p=0.95,
)
content = response.choices[0].message.content

print(30 * "=" + "BOT 1" + 30 * "=")
print(content)
print("\n\n")
# The model could give the following answer:
# ```L'orateur le plus inspirant est le président.
# Il est plus inspirant parce qu'il parle de ses expériences personnelles
# et de son optimisme pour l'avenir du pays.
# Il est différent de l'autre orateur car il ne parle pas de la météo,
# mais plutôt de ses interactions avec les gens et de son rôle en tant que président.```

messages = [
    user_msg,
    AssistantMessage(content=content).to_openai(),
    UserMessage(content="Ok, now please summarize the content of the first audio.").to_openai()
]
print(30 * "=" + "USER 2" + 30 * "=")
print(messages[-1]["content"])
print("\n\n")

response = client.chat.completions.create(
    model=model,
    messages=messages,
    temperature=0.2,
    top_p=0.95,
)
content = response.choices[0].message.content
print(30 * "=" + "BOT 2" + 30 * "=")
print(content)

Transcription

Voxtral-Small-24B-2507 has powerful transcription capabilities!

Make sure that your client has mistral-common with audio installed:

pip install --upgrade mistral_common\[audio\]

Python snippet

from mistral_common.protocol.transcription.request import TranscriptionRequest
from mistral_common.protocol.instruct.messages import RawAudio
from mistral_common.audio import Audio
from huggingface_hub import hf_hub_download

from openai import OpenAI

# Modify OpenAI's API key and API base to use vLLM's API server.
openai_api_key = "EMPTY"
openai_api_base = "http://<your-server-host>:8000/v1"

client = OpenAI(
    api_key=openai_api_key,
    base_url=openai_api_base,
)

models = client.models.list()
model = models.data[0].id

obama_file = hf_hub_download("patrickvonplaten/audio_samples", "obama.mp3", repo_type="dataset")
audio = Audio.from_file(obama_file, strict=False)

audio = RawAudio.from_audio(audio)
req = TranscriptionRequest(model=model, audio=audio, language="en", temperature=0.0).to_openai(exclude=("top_p", "seed"))

response = client.audio.transcriptions.create(**req)
print(response)

Function Calling

Voxtral has some experimental function calling support. You can try as shown below.

Make sure that your client has mistral-common with audio installed:

pip install --upgrade mistral_common\[audio\]

Python snippet

from mistral_common.protocol.instruct.messages import AudioChunk, UserMessage, TextChunk
from mistral_common.protocol.transcription.request import TranscriptionRequest
from mistral_common.protocol.instruct.tool_calls import Function, Tool

from mistral_common.audio import Audio
from huggingface_hub import hf_hub_download

from openai import OpenAI

# Modify OpenAI's API key and API base to use vLLM's API server.
openai_api_key = "EMPTY"
openai_api_base = "http://<your-server-host>:8000/v1"

client = OpenAI(
    api_key=openai_api_key,
    base_url=openai_api_base,
)

models = client.models.list()
model = models.data[0].id

tool = Tool(
    function=Function(
        name="get_current_weather",
        description="Get the current weather",
        parameters={
            "type": "object",
            "properties": {
                "location": {
                    "type": "string",
                    "description": "The city and state, e.g. San Francisco, CA",
                },
                "format": {
                    "type": "string",
                    "enum": ["celsius", "fahrenheit"],
                    "description": "The temperature unit to use. Infer this from the user's location.",
                },
            },
            "required": ["location", "format"],
        },
    )
)
tools = [tool.to_openai()]


weather_like = hf_hub_download("patrickvonplaten/audio_samples", "fn_calling.wav", repo_type="dataset")

def file_to_chunk(file: str) -> AudioChunk:
    audio = Audio.from_file(file, strict=False)
    return AudioChunk.from_audio(audio)

audio_chunk = file_to_chunk(weather_like)

print(30 * "=" + "Transcription" + 30 * "=")
req = TranscriptionRequest(model=model, audio=audio_chunk.input_audio, language="en", temperature=0.0).to_openai(exclude=("top_p", "seed"))
response = client.audio.transcriptions.create(**req)
print(response.text) # How is the weather in Madrid at the moment?
print("\n")


print(30 * "=" + "Function calling" + 30 * "=")
audio_chunk = file_to_chunk(weather_like)
user_msg = UserMessage(content=[audio_chunk]).to_openai()
response = client.chat.completions.create(
    model=model,
    messages=[user_msg],
    temperature=0.2,
    top_p=0.95,
    tools=[tool.to_openai()]
)
print(30 * "=" + "BOT 1" + 30 * "=")
print(response.choices[0].message.tool_calls)
print("\n\n")

Transformers 🤗

Starting with transformers >= 4.54.0 and above, you can run Voxtral natively!

Install Transformers:

pip install -U transformers

Make sure to have mistral-common >= 1.8.1 installed with audio dependencies:

pip install --upgrade "mistral-common[audio]"

Audio Instruct

➡️ multi-audio + text instruction

from transformers import VoxtralForConditionalGeneration, AutoProcessor
import torch

device = "cuda"
repo_id = "mistralai/Voxtral-Small-24B-2507"

processor = AutoProcessor.from_pretrained(repo_id)
model = VoxtralForConditionalGeneration.from_pretrained(repo_id, torch_dtype=torch.bfloat16, device_map=device)

conversation = [
    {
        "role": "user",
        "content": [
            {
                "type": "audio",
                "path": "https://huggingface.co/datasets/hf-internal-testing/dummy-audio-samples/resolve/main/mary_had_lamb.mp3",
            },
            {
                "type": "audio",
                "path": "https://huggingface.co/datasets/hf-internal-testing/dummy-audio-samples/resolve/main/winning_call.mp3",
            },
            {"type": "text", "text": "What sport and what nursery rhyme are referenced?"},
        ],
    }
]

inputs = processor.apply_chat_template(conversation)
inputs = inputs.to(device, dtype=torch.bfloat16)

outputs = model.generate(**inputs, max_new_tokens=500)
decoded_outputs = processor.batch_decode(outputs[:, inputs.input_ids.shape[1]:], skip_special_tokens=True)

print("\nGenerated response:")
print("=" * 80)
print(decoded_outputs[0])
print("=" * 80)

➡️ multi-turn

from transformers import VoxtralForConditionalGeneration, AutoProcessor
import torch

device = "cuda"
repo_id = "mistralai/Voxtral-Small-24B-2507"

processor = AutoProcessor.from_pretrained(repo_id)
model = VoxtralForConditionalGeneration.from_pretrained(repo_id, torch_dtype=torch.bfloat16, device_map=device)

conversation = [
    {
        "role": "user",
        "content": [
            {
                "type": "audio",
                "path": "https://huggingface.co/datasets/hf-internal-testing/dummy-audio-samples/resolve/main/obama.mp3",
            },
            {
                "type": "audio",
                "path": "https://huggingface.co/datasets/hf-internal-testing/dummy-audio-samples/resolve/main/bcn_weather.mp3",
            },
            {"type": "text", "text": "Describe briefly what you can hear."},
        ],
    },
    {
        "role": "assistant",
        "content": "The audio begins with the speaker delivering a farewell address in Chicago, reflecting on his eight years as president and expressing gratitude to the American people. The audio then transitions to a weather report, stating that it was 35 degrees in Barcelona the previous day, but the temperature would drop to minus 20 degrees the following day.",
    },
    {
        "role": "user",
        "content": [
            {
                "type": "audio",
                "path": "https://huggingface.co/datasets/hf-internal-testing/dummy-audio-samples/resolve/main/winning_call.mp3",
            },
            {"type": "text", "text": "Ok, now compare this new audio with the previous one."},
        ],
    },
]

inputs = processor.apply_chat_template(conversation)
inputs = inputs.to(device, dtype=torch.bfloat16)

outputs = model.generate(**inputs, max_new_tokens=500)
decoded_outputs = processor.batch_decode(outputs[:, inputs.input_ids.shape[1]:], skip_special_tokens=True)

print("\nGenerated response:")
print("=" * 80)
print(decoded_outputs[0])
print("=" * 80)

➡️ text only

from transformers import VoxtralForConditionalGeneration, AutoProcessor
import torch

device = "cuda"
repo_id = "mistralai/Voxtral-Small-24B-2507"

processor = AutoProcessor.from_pretrained(repo_id)
model = VoxtralForConditionalGeneration.from_pretrained(repo_id, torch_dtype=torch.bfloat16, device_map=device)

conversation = [
    {
        "role": "user",
        "content": [
            {
                "type": "text",
                "text": "Why should AI models be open-sourced?",
            },
        ],
    }
]

inputs = processor.apply_chat_template(conversation)
inputs = inputs.to(device, dtype=torch.bfloat16)

outputs = model.generate(**inputs, max_new_tokens=500)
decoded_outputs = processor.batch_decode(outputs[:, inputs.input_ids.shape[1]:], skip_special_tokens=True)

print("\nGenerated response:")
print("=" * 80)
print(decoded_outputs[0])
print("=" * 80)

➡️ audio only

from transformers import VoxtralForConditionalGeneration, AutoProcessor
import torch

device = "cuda"
repo_id = "mistralai/Voxtral-Small-24B-2507"

processor = AutoProcessor.from_pretrained(repo_id)
model = VoxtralForConditionalGeneration.from_pretrained(repo_id, torch_dtype=torch.bfloat16, device_map=device)

conversation = [
    {
        "role": "user",
        "content": [
            {
                "type": "audio",
                "path": "https://huggingface.co/datasets/hf-internal-testing/dummy-audio-samples/resolve/main/winning_call.mp3",
            },
        ],
    }
]

inputs = processor.apply_chat_template(conversation)
inputs = inputs.to(device, dtype=torch.bfloat16)

outputs = model.generate(**inputs, max_new_tokens=500)
decoded_outputs = processor.batch_decode(outputs[:, inputs.input_ids.shape[1]:], skip_special_tokens=True)

print("\nGenerated response:")
print("=" * 80)
print(decoded_outputs[0])
print("=" * 80)

➡️ batched inference

from transformers import VoxtralForConditionalGeneration, AutoProcessor
import torch

device = "cuda"
repo_id = "mistralai/Voxtral-Small-24B-2507"

processor = AutoProcessor.from_pretrained(repo_id)
model = VoxtralForConditionalGeneration.from_pretrained(repo_id, torch_dtype=torch.bfloat16, device_map=device)

conversations = [
    [
        {
            "role": "user",
            "content": [
                {
                    "type": "audio",
                    "path": "https://huggingface.co/datasets/hf-internal-testing/dummy-audio-samples/resolve/main/obama.mp3",
                },
                {
                    "type": "audio",
                    "path": "https://huggingface.co/datasets/hf-internal-testing/dummy-audio-samples/resolve/main/bcn_weather.mp3",
                },
                {
                    "type": "text",
                    "text": "Who's speaking in the speach and what city's weather is being discussed?",
                },
            ],
        }
    ],
    [
        {
            "role": "user",
            "content": [
                {
                    "type": "audio",
                    "path": "https://huggingface.co/datasets/hf-internal-testing/dummy-audio-samples/resolve/main/winning_call.mp3",
                },
                {"type": "text", "text": "What can you tell me about this audio?"},
            ],
        }
    ],
]

inputs = processor.apply_chat_template(conversations)
inputs = inputs.to(device, dtype=torch.bfloat16)

outputs = model.generate(**inputs, max_new_tokens=500)
decoded_outputs = processor.batch_decode(outputs[:, inputs.input_ids.shape[1]:], skip_special_tokens=True)

print("\nGenerated responses:")
print("=" * 80)
for decoded_output in decoded_outputs:
    print(decoded_output)
    print("=" * 80)

Transcription

➡️ transcribe

from transformers import VoxtralForConditionalGeneration, AutoProcessor
import torch

device = "cuda"
repo_id = "mistralai/Voxtral-Small-24B-2507"

processor = AutoProcessor.from_pretrained(repo_id)
model = VoxtralForConditionalGeneration.from_pretrained(repo_id, torch_dtype=torch.bfloat16, device_map=device)

inputs = processor.apply_transcription_request(language="en", audio="https://huggingface.co/datasets/hf-internal-testing/dummy-audio-samples/resolve/main/obama.mp3", model_id=repo_id)
inputs = inputs.to(device, dtype=torch.bfloat16)

outputs = model.generate(**inputs, max_new_tokens=500)
decoded_outputs = processor.batch_decode(outputs[:, inputs.input_ids.shape[1]:], skip_special_tokens=True)

print("\nGenerated responses:")
print("=" * 80)
for decoded_output in decoded_outputs:
    print(decoded_output)
    print("=" * 80)