---
title: GritLM/GritLM-7B
description: > GritLM is a generative representational instruction tuned language model. It unifies text representation (embedding) and text generation i. Mistral, 7.2B parameters.
canonical_url: https://superlinked.com/models/gritlm-gritlm-7b
last_updated: 2026-06-05
---

# GritLM/GritLM-7B

> GritLM is a generative representational instruction tuned language model. It unifies text representation (embedding) and text generation into a single model achieving state-of-the-art performance on both types of tasks.

Source: [GritLM/GritLM-7B on HuggingFace](https://huggingface.co/GritLM/GritLM-7B)
Base model: [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1)

## Overview

| Field | Value |
|-------|-------|
| Architecture | Mistral |
| Parameters | 7.2B |
| Tasks | Encode |
| Outputs | Dense |
| Dimensions | Dense: 4,096 |
| Max sequence length | 4,096 tokens |
| License | apache-2.0 |
| Inputs | text |

## Benchmarks

### NFCorpus

Domain: medical · Task: retrieval · Language: en

Biomedical literature search from NutritionFacts.org

Corpus: 3,593 · Queries: 323

**Quality:** ndcg at 10: 0.3972 · map at 10: 0.1531 · mrr at 10: 0.6139

**Performance (L4 b1 c16):** Corpus 1.7K tok/s · Corpus p50 2.7s · Query 312 tok/s · Query p50 206.2ms

[Reference](https://www.cl.uni-heidelberg.de/statnlpgroup/nfcorpus/)

### NanoFiQA2018Retrieval

Domain: finance · Task: retrieval · Language: en

Smaller subset of the FiQA financial QA dataset

**Quality:** ndcg at 10: 0.6289 · map at 10: 0.5506 · mrr at 10: 0.6275

**Performance (L4 b1 c16):** Corpus 1.1K tok/s · Corpus p50 1.6s · Query 556 tok/s · Query p50 196.8ms

[Reference](https://sites.google.com/view/fiqa/)