MODEL PROFILE

Voxtral-Mini-4B-Realtime-2602

Mistral AI · Voxtral-Mini-4B-Realtime-2602

Voxtral Mini 4B Realtime 2602 is a **multilingual, realtime speech-transcription model** and among the first open-source solutions to achieve accuracy comparable to offline systems with a delay of ** = 3600 / 0.8 = 45000`. In theory, you should be able to record with no limit; in practice, pre-allocations of RoPE parameters among other things limits `--max-model-len`. For the best user experience, we recommend to simply instantiate vLLM with the default parameters which will automatically set a maximum model length of 131072 (~ca. 3h). - We strongly recommend using websockets to set up audio streaming sessions. For more info on how to do so, check Usage. - We recommend using a delay of 480ms as we found it to be the sweet spot of performance and low latency. If, however, you want to adapt the delay, you can change the `"transcription_delay_ms": 480` parameter in the tekken.json file to any multiple of 80ms between 80 and 1200, as well as 2400 as a standalone value. We compare Voxtral Mini 4B Realtime to similar models - both offline models and realtime. Voxtral Mini 4B Realtime is competitive to leading offline models and shows significant gains over existing open-source realtime solutions. > [!Warning] > Due to its novel architecture, Voxtral Realtime is currently only support in vLLM. We very much welcome community contributions > to add the architecture to Transformers and Llama.cpp.

Text generationlicense:apache-2.0
Context windowNot disclosed
Input → OutputAudio → Text
Model weightsOpen weights
Profile updated2026/11/16 05:55:27
OVERVIEW

Model overview

Voxtral Mini 4B Realtime 2602 is a **multilingual, realtime speech-transcription model** and among the first open-source solutions to achieve accuracy comparable to offline systems with a delay of ** = 3600 / 0.8 = 45000`. In theory, you should be able to record with no limit; in practice, pre-allocations of RoPE parameters among other things limits `--max-model-len`. For the best user experience, we recommend to simply instantiate vLLM with the default parameters which will automatically set a maximum model length of 131072 (~ca. 3h). - We strongly recommend using websockets to set up audio streaming sessions. For more info on how to do so, check Usage. - We recommend using a delay of 480ms as we found it to be the sweet spot of performance and low latency. If, however, you want to adapt the delay, you can change the `"transcription_delay_ms": 480` parameter in the tekken.json file to any multiple of 80ms between 80 and 1200, as well as 2400 as a standalone value. We compare Voxtral Mini 4B Realtime to similar models - both offline models and realtime. Voxtral Mini 4B Realtime is competitive to leading offline models and shows significant gains over existing open-source realtime solutions. > [!Warning] > Due to its novel architecture, Voxtral Realtime is currently only support in vLLM. We very much welcome community contributions > to add the architecture to Transformers and Llama.cpp.

The summary below is compiled from sources by an AI model, not our own testing. Verify selection decisions against original sources and your workload.

Key features

  • Multilingual real-time speech transcription with sub-second latency (optimized at 480ms delay)
  • Open-source model under Apache-2.0 license, designed for streaming audio input
  • Supports up to ~3 hours of continuous audio via vLLM's default max-model-len of 131072 tokens

Limitations & considerations

  • Currently only compatible with vLLM inference engine — no support in Transformers or Llama.cpp (as of documentation)
  • Transcription delay must be set to multiples of 80ms (80–1200ms) or 2400ms; no arbitrary values allowed
  • Maximum model length is constrained by RoPE pre-allocation limits, making truly unlimited recording impractical despite theoretical capacity
PRICING & COST

Source prices & cost analysis

Prices vary by channel, region and configuration. Channel quotes are listed independently and are not treated as a single official provider price.

Source / channelInputOutputBilling unitCollected at
No verified prices yet. Check the source channel.

Estimate monthly base token costs

—

Estimates exclude cache discounts, tiered pricing, multimodal surcharges, web search, taxes and additional reasoning output.

INDEPENDENT BENCHMARKS

Public benchmarks & run configuration

暂无已采集的可核验公开评分。模型未上榜不代表能力较低。

View full indexed rankings →

Frequently asked questions

What is the recommended transcription delay and why?

The recommended delay is 480ms, identified as the optimal balance between low latency and transcription performance.

Which inference engines support Voxtral Mini 4B Realtime?

It is currently supported only in vLLM; support for Hugging Face Transformers and Llama.cpp is not available but welcomed via community contributions.

How long of an audio stream can it process?

With vLLM's default configuration, it supports up to ~3 hours (~131,072 tokens), though practical limits arise from RoPE parameter pre-allocation.

Do open weights permit free commercial use?

Not necessarily. Check the version license, commercial conditions and usage policies. Self-hosting incurs compute and maintenance costs.

Are these real-time billing prices?

These are timestamped source snapshots, not live billing prices. Check the channel for current versions, regions, discounts and billing methods.

Sources & traceability

许可证:apache-2.0。模型介绍与第三方评测均以原来源的具体版本和方法为准。