# Mistral 7B Instruct v0.1

Mistral AI's first 7B instruction-tuned model, built on an efficient transformer architecture with an 8k context window for general-purpose chat and generation.

## about

Mistral AI's debut model introduced sliding window attention with a 4,096-token window that stacks across layers to reach an effective 32K-token span, a novel approach [released via a torrent link on X](https://mistral.ai/news/announcing-mistral-7b) with no paper or blog post. At 7.24B parameters it outperformed Llama 2 13B on every benchmark, and became the most popular base for community fine-tuning in late 2023.

- License: Apache-2.0  
- Context window (in thousands): 8192

## Use cases for Mistral 7B Instruct v0.1

1. **Efficient chatbot deployment:** Sliding window attention delivers 32K effective context at 7B-parameter inference costs, enabling responsive assistants on single-GPU setups.
2. **Community model fine-tuning:** As the most-forked open base model of late 2023, its architecture and Apache 2.0 license make it the standard starting point for domain-specific instruction tuning.
3. **Multilingual content generation:** Its grouped-query attention and broad pretraining data support generation across European languages with lower latency than comparably capable dense models.

## Quality

- Arena Elo: 1008  
- MMLU: 55.4  
- MT Bench: 6.84

Mistral 7B Instruct v0.1 scores 56.3% on MMLU and 6.84 on MT-Bench, placing it below Gemma 7B IT (64.3% MMLU) on knowledge but introducing sliding window attention as an architectural innovation at the 7B scale. Despite the lower MMLU score, its efficient inference and Apache 2.0 license made it the most-forked open base model of late 2023.

**Competitor Scores:**  
**Gemma 7B IT:** 1038  
**Llama 2 Chat 7B:** 1037  
**Nous Hermes 2 Mistral 7B:** 1010  
**Mistral 7B Instruct v0.1:** 1008  
**Gemma 2B IT:** 990

## pricing

The cost per 1,000 tokens for running the model with Telnyx Inference is $0.0003. For instance, analyzing 1,000,000 customer chats, assuming each chat is 1,000 tokens, would cost $300.

## What's Twitter saying?

- Developers report **poor inference speed** with Mistral-7B-Instruct-v0.1 on RTX 3090 Ti GPUs, taking ~60s for basic questions despite 100% utilization, slower than Ollama on M1 MacBook Air.
- Fine-tuned versions are **mind-blowing for products**, running flawlessly on RTX 3090 and serviceably on M1 MBP, praised on Hacker News.
- Users call it **amazing for local runs** on personal computers, better than early LLaMA models despite some limitations.

## faqs

### What is Mistral 7B Instruct v0.1?

Mistral 7B Instruct v0.1 is the first instruction-tuned variant of [Mistral AI's 7.3 billion parameter base model](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.1), fine-tuned for conversational and instruction-following tasks. It was released in September 2023 under the Apache 2.0 license.

### Is Mistral 7B better than GPT-3?

Mistral 7B outperforms GPT-3 on most standard benchmarks despite being a fraction of the size. Its [efficient architecture](/content/llm-library/mistral-7b-instruct-v0-1/index.html) with grouped-query attention enables competitive reasoning and code generation at significantly lower compute costs.

### What is the Mistral 7B model?

Mistral 7B is a 7.3 billion parameter language model developed by Mistral AI, designed for efficiency without sacrificing capability. The [research paper](https://arxiv.org/abs/2310.06825) demonstrated that it outperforms Llama 2 13B on all benchmarks and matches Llama 1 34B on several.

### Can you run Mistral AI locally?

Yes, Mistral 7B runs on consumer hardware with as little as 8GB of VRAM using quantized formats. Tools like Ollama and llama.cpp make it straightforward to deploy [Mistral 7B Instruct](/content/llm-library/mistral-7b-instruct-v0-1/index.html) locally for development and testing.

### Can I use Mistral 7B for free?

Mistral 7B is released under the Apache 2.0 license, making it free for both personal and commercial use with no restrictions. You can download the weights from [Hugging Face](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.1) or access it through hosted inference providers.

### Is Mistral as good as ChatGPT?

Mistral 7B is not a direct competitor to ChatGPT's underlying GPT-4 models, which are significantly larger. For resource-constrained deployments and specific tasks like [code assistance and summarization](/content/llm-library/hermes-2-pro-mistral-7b/index.html), Mistral 7B provides strong results at a fraction of the cost.

### Why is Mistral 7B so good?

Mistral 7B achieves outsized performance through sliding window attention for long contexts and grouped-query attention for fast inference. These architectural choices, documented in the [original research](https://arxiv.org/abs/2310.06825), let a 7B model compete with models up to 34B parameters on key benchmarks.

---  
Mistral 7B Instruct v0.1—High quality, low latency LLM.