Model reference · open weights

Borealis

Audio Vikhrmodels Speech→text 1 build Open weights 74k dl/mo

Borealis is an open-weight audio or speech model from Vikhrmodels. Borealis (BF16) weighs 2.3 GB; the smallest configuration that runs it is RTX 3060 12 GB.

For self-hosting, compare this build's requirements below with the GPU you plan to rent.

  • Borealis is an automatic speech recognition model developed by Vikhrmodels for the Russian language.
  • It is an audio large language model trained on approximately 7,000 hours of audio data and features punctuation support in its output.
  • The model is released under the Apache 2.0 license.

Summary of the Vikhrmodels/Borealis model card, 2026-10-01

What it is

Released byVikhrmodels
Released2025-09-11
VRAM2.3 GB for the weights

What it runs on

Memory and cards for Borealis (BF16)

2.3 GBweights, file size
1.6 GBruntime overhead
CardOne streamMemory
RTX 3060 12 GBfits11.6 GB
RTX 4060 Ti 16 GBfits15.4 GB
RTX 3090 24 GBfits23.4 GB
RTX 4090 24 GBfits23.4 GB
RTX 5090 32 GBfits31.0 GB
L40S 48 GBfits44.0 GB
A100 80 GBfits78.2 GB
H100 80 GBfits78.1 GB
RTX PRO 6000 Blackwell 96 GBfits93.8 GB
DGX Spark (GB10) 128 GB unifiedfits107 GB
H200 141 GBfits138 GB
B200 180 GBfits176 GB

From the model card

What Vikhrmodels says about Borealis

Read the model card

Borealis

Описание

Borealis - это наша первая audio llm c ASR для русского языка. В этом репозитории представлен чекпоинт, который видел примерно 7000 часов аудио на русском языке. Важным отличием от других моделей является то, что тут есть поддержка пунктуации в распознанных аудио. Архитектура во многом вдохновлена Voxtral, но отличается в некоторых моментах.

Использование

from transformers import AutoModelForCausalLM, AutoTokenizer, AutoFeatureExtractor
import torch
import librosa

model = AutoModelForCausalLM.from_pretrained("Vikhrmodels/Borealis", trust_remote_code=True)
tokenizer = AutoTokenizer.from_pretrained("Vikhrmodels/Borealis")
extractor = AutoFeatureExtractor.from_pretrained("Vikhrmodels/Borealis")

generation_params = {
        "max_new_tokens": 350,
        "do_sample": True,
        "top_p": 0.9,
        "top_k": 50,
        "temperature": 0.2,
 }

model.eval()
model.to("cuda")

waveform, sr = librosa.load("path/to/your/audio.wav", sr=16_000)

proc = extractor(
        waveform,
        sampling_rate=sr,
        padding="max_length",
        max_length=480_000,
        return_attention_mask=True,
        return_tensors="pt",
    )

mel = proc.input_features.squeeze(0).to("cuda")
att_mask = proc.attention_mask.squeeze(0).to("cuda")

with torch.inference_mode():
    transcript = model.generate(mel=mel, att_mask=att_mask, **generation_params)

print(transcript)

Метрики модели

Ниже представлены замеры Borealis на фоне остальных открытых моделей, который поддерживают русский язык. Бенчмарк мы скоро выложим в открытый доступ

МодельСредний WERRuLSCV 22.0BooksSpeakSova
Borealis6.33%6.39%2.67%5.28%1.95%15.37%
GigaAM-ASR-V2-RNNT5.85%5.24%2.85%8.06%3.08%10.01%
openai/whisper-large-v310.74%11.62%7.51%12.19%2.74%19.65%
bond005/whisper-podlodka-turbo9.38%11.91%6.36%8.96%3.14%16.55%
openai/whisper-large-v3-turbo11.30%11.88%8.17%13.29%2.80%20.37%
bond005/whisper-large-v3-ru-podlodka10.76%10.24%7.80%10.31%3.23%22.21%
nvidia/canary-1b-v213.52%20.16%9.12%11.45%4.97%21.89%
VOSK-model-ru-0.4211.30%12.06%11.87%10.80%2.61%19.15%
GigaAM-ASR-V2-CTC6.45%5.26%3.42%7.72%3.01%12.86%

Авторы

  • Илья Кулешов, Vikhr Team
  • Александр Николич, Vikhr Team

Quoted from the model card on Hugging Face. The full card is behind the Hugging Face link above.

© 2026 AxForge · EU-hosted AI infrastructure Pricing Docs Trust Privacy Terms