Mistral AI Introduces 'Le Chat' Mobile App, And A 1,000 Words Per Second Response Speed

How fast it too fast? To some people, they just want more than they can handle.

It's said that the average reading speed for most people is 200–250 words per minute (WPM) with normal comprehension. However, trained speed readers can reach much higher speeds. World-class spread readers can read around 2,000 WPM. Some individuals claim to read at 25,000+ WPM, but these often involve skimming with minimal comprehension.

So, wow about 1,000 tokens per second?

This number that can roughly be translated to 750 words per second in Natural Language Processing for English text, is around 45,000 VPM, is about twice as much as the fastest speed reader.

Mistral AI can answer users' queries that fast.

Faster than any human eyes can read.

Paris-based AI startup Mistral AI has introduced a series of major updates to its AI assistant, Le Chat, including new features, mobile applications, and a revised pricing structure.

Regarded as an underdog when compared to OpenAI, Google, Microsoft and some others, Mistral AI entered the competition following the hype of ChatGPT.

To better complete, Mistral AI that has upgraded its Le Chat's web platform, has also launched dedicated iOS and Android apps, in order to make the AI chatbot a lot more accessible to mobile users.

Since Large Language Models-powered chatbots do pretty much the same thing and provide similar features, Mistral AI makes Le Chat boast something others cannot match: speed.

The key update here, is the introduction of 'Flash Answers,' which is a feature available to all user tiers.

What it does, is allowing the chatbot to deliver responses at speeds of up to 1,000 tokens per second, making it one of the fastest AI assistants on the market.

"Powered by the highest-performing, lowest-latency Mistral models and the fastest inference engines on the planet, le Chat reasons, reflects, and responds faster than any other chat assistant, up to ~1000 words / sec," said Mistral AI in the announcement.

However, this speed advantage applies only to text-based responses—image generation may take longer.

Le Chat now includes a code interpreter, enabling users to execute code, analyze data, create visualizations, and run simulations directly within the assistant.

This feature supports Python and allows users to upload CSV files for quick data insights. It simplifies tasks such as mathematical computations, data plotting, and scientific analysis.

For comparison, Anthropic’s Claude AI introduced a similar capability last year, utilizing JavaScript for complex calculations and visualizations.

Mistral AI has enhanced Le Chat’s image generation capabilities using Flux Ultra, a cutting-edge model from Black Forest Labs Inc.

According to Mistral, Flux Ultra surpasses image generation quality from competitors like OpenAI and Grok, offering users both photorealistic and highly imaginative visuals.

Mistral previously integrated web search with citations, allowing users to access accurate, real-time information—similar to AI platforms like ChatGPT, Gemini, and Perplexity.

More recently, the company announced a partnership with Agence France-Presse (AFP), a leading global news agency. This integration enriches Le Chat's responses with up-to-date, fact-based news coverage.

"There's also some elevated privacy as well where you can fully opt out of any data recording, which is very useful for professionals," co-founder and CEO Arthur Mensch said in an interview.

At this time, Mistral is one of Europe's most-funded AI startups.