---
title: "AI models released in July 2024"
type: "release-archive"
canonical_url: "https://theaiengineering.com/releases/2024/07"
markdown_url: "https://theaiengineering.com/releases/2024/07.md"
author: "The AI Engineering Editors"
publisher: "The AI Engineering"
updated: "2024-07-31"
citation: "The AI Engineering Editors. \"AI models released in July 2024.\" The AI Engineering, updated Jul 31, 2024. https://theaiengineering.com/releases/2024/07"
license: "CC BY 4.0"
license_url: "https://creativecommons.org/licenses/by/4.0/"
data_csv: "https://theaiengineering.com/releases.csv"
data_json: "https://theaiengineering.com/releases.json"
---

# AI models released in July 2024

20 releases from 12 labs, 16 with open weights. Dates are the labs' official announcement dates.

## July 2024

### Gemma 2 2B

Gemma 2 2B is a new 2B-parameter version of Gemma built on the Gemma 2 architecture, designed to run efficiently on hardware from phones and laptops to cloud deployments on Vertex AI and GKE.

- Date: 2024-07-31
- Lab: Google
- Type: Small model
- Weights: Open weights
- Modalities: text → text
- Parameters: 2B
- License: Gemma Terms of Use
- Available at launch: Open weights, API, Cloud platforms
- Permalink: https://theaiengineering.com/releases/2024/07#gemma-2-2b

Highlights:

- Weights on Kaggle, Hugging Face and Vertex AI Model Garden
- Can be tried in Google AI Studio
- Released under Gemma terms for research and commercial use

Source: [Smaller, Safer, More Transparent: Advancing Responsible AI with Gemma](https://developers.googleblog.com/en/smaller-safer-more-transparent-advancing-responsible-ai-with-gemma/)

### ShieldGemma

ShieldGemma is a set of safety classifiers built on Gemma 2 that filter the inputs and outputs of AI models. It detects hate speech, harassment, sexually explicit content and dangerous content.

- Date: 2024-07-31
- Lab: Google
- Type: LLM
- Weights: Open weights
- Modalities: text → text
- Parameters: 2B, 9B, 27B
- License: Gemma Terms of Use
- Available at launch: Open weights
- Permalink: https://theaiengineering.com/releases/2024/07#shieldgemma

Highlights:

- 2B suits online classification; 9B and 27B target offline use
- Released alongside Gemma 2 2B and the Gemma Scope interpretability tools

Source: [Smaller, Safer, More Transparent: Advancing Responsible AI with Gemma](https://developers.googleblog.com/en/smaller-safer-more-transparent-advancing-responsible-ai-with-gemma/)

### Midjourney V6.1

Point release of V6 with more coherent bodies and objects, better image quality and small details, new 2x upscalers, improved text accuracy and a new personalization model, made default at launch.

- Date: 2024-07-30
- Lab: Midjourney
- Type: Image
- Weights: Proprietary
- Modalities: text, image → image
- Available at launch: Consumer app
- Permalink: https://theaiengineering.com/releases/2024/07#midjourney-v6-1

Highlights:

- About 25% faster for standard image jobs
- Default model until June 16, 2025
- Personalization code versioning

Source: [Version 6.1](https://updates.midjourney.com/version-6-1/)

### SAM 2

SAM 2 is Meta's next-generation Segment Anything Model, a unified model for real-time promptable object segmentation in both images and videos.

- Date: 2024-07-29
- Lab: Meta
- Type: Multimodal
- Weights: Open weights
- Modalities: image, video → image, video
- License: Apache-2.0
- Available at launch: Open weights
- Permalink: https://theaiengineering.com/releases/2024/07#meta-sam-2

Highlights:

- Runs at about 44 frames per second
- Released with the SA-V dataset of about 51,000 videos under CC BY 4.0
- Treats an image as a single-frame video so one architecture handles both

Source: [Introducing SAM 2: The next generation of Meta Segment Anything Model for videos and images](https://ai.meta.com/blog/segment-anything-2/)

### DeepSeek-Coder-V2-0724

DeepSeek-Coder-V2-0724 is an updated DeepSeek-Coder-V2 checkpoint with improved general capabilities through alignment optimization. It upgraded the deepseek-coder API model and was later merged into DeepSeek-V2.5.

- Date: 2024-07-24
- Lab: DeepSeek
- Type: Coding
- Weights: Open weights
- Modalities: text → text
- Context window: 128,000 tokens
- Parameters: MoE: 236B total, 21B active
- License: DeepSeek Model License
- Available at launch: API, Open weights
- Permalink: https://theaiengineering.com/releases/2024/07#deepseek-coder-v2-0724

Highlights:

- Upgraded the deepseek-coder API model on July 24, 2024
- Weights published on Hugging Face as DeepSeek-Coder-V2-Instruct-0724
- One of the two models merged to create DeepSeek-V2.5

Source: [Change Log | DeepSeek API Docs](https://api-docs.deepseek.com/updates)

### Mistral Large 2

Second-generation flagship with 123B parameters and a 128K-token context window, designed for single-node inference and released with weights under the Mistral Research License.

- Date: 2024-07-24
- Lab: Mistral AI
- Type: LLM
- Weights: Open weights
- Modalities: text → text
- Context window: 128,000 tokens
- Parameters: 123B
- License: Mistral Research License
- Available at launch: Open weights, API
- Permalink: https://theaiengineering.com/releases/2024/07#mistral-large-2

Highlights:

- Supports dozens of languages and 80+ coding languages
- Commercial self-deployment requires a Mistral Commercial License
- API model ID mistral-large-2407

Source: [Large Enough | Mistral AI](https://mistral.ai/news/mistral-large-2407)

### Llama 3.1

Llama 3.1 introduced the 405B model, which Meta described as the largest openly available foundation model, plus upgraded multilingual 8B and 70B models, all with a 128K context window.

- Date: 2024-07-23
- Lab: Meta
- Type: LLM
- Weights: Open weights
- Modalities: text → text
- Context window: 128,000 tokens
- Parameters: 8B, 70B, 405B
- License: Llama 3.1 Community License
- Available at launch: Open weights, Cloud platforms
- Permalink: https://theaiengineering.com/releases/2024/07#llama-3-1

Highlights:

- Supports eight languages
- License changed to allow using Llama outputs to improve other models
- 405B trained on over 15 trillion tokens
- Over 25 partners, including AWS, Azure and Google Cloud, offered services on day one

Source: [Introducing Llama 3.1: Our most capable models to date](https://ai.meta.com/blog/meta-llama-3-1/)

### Llama Guard 3

Llama Guard 3 is a multilingual content safety classifier fine-tuned from Llama 3.1 8B, released with Llama 3.1 and optimized for search and code interpreter tool calls.

- Date: 2024-07-23
- Lab: Meta
- Type: LLM
- Weights: Open weights
- Modalities: text → text
- Parameters: 8B
- License: Llama 3.1 Community License
- Available at launch: Open weights
- Permalink: https://theaiengineering.com/releases/2024/07#llama-guard-3

Highlights:

- Predicts safety labels across 14 categories, adding Code Interpreter Abuse
- Supports eight languages

Source: [Introducing Llama 3.1: Our most capable models to date](https://ai.meta.com/blog/meta-llama-3-1/)

### Prompt Guard

Prompt Guard is a small classifier released with Llama 3.1 that labels inputs as benign, prompt injection or jailbreak, built to filter prompt attacks in LLM applications.

- Date: 2024-07-23
- Lab: Meta
- Type: Small model
- Weights: Open weights
- Modalities: text → text
- Context window: 512 tokens
- Parameters: 86M
- License: Llama 3.1 Community License
- Available at launch: Open weights
- Permalink: https://theaiengineering.com/releases/2024/07#prompt-guard

Highlights:

- Fine-tuned from mDeBERTa-v3-base with 86M backbone parameters
- 512-token context window; longer inputs are split into segments

Source: [Introducing Llama 3.1: Our most capable models to date](https://ai.meta.com/blog/meta-llama-3-1/)

### Rerank 3 Nimble

Faster member of the Rerank 3 series, about three times faster than Rerank 3 while keeping high accuracy, offered in English and multilingual versions.

- Date: 2024-07-23
- Lab: Cohere
- Type: Reranker
- Weights: Proprietary
- Modalities: text → score
- Available at launch: API
- Permalink: https://theaiengineering.com/releases/2024/07#cohere-rerank-3-nimble

Highlights:

- Multilingual version covers 100+ languages
- Can rerank documents tens of thousands of tokens long
- Handles emails, JSON and tabular data

Source: [Introducing Rerank 3 Nimble: Faster Reranking for Enterprise Search and Retrieval-Augmented Generation (RAG) Systems](https://cohere.com/blog/rerank-3-nimble)

### Eleven Turbo v2.5

Low-latency multilingual model that made Hindi, French, Spanish, Mandarin and 27 other languages three times faster and English 25% faster for conversational AI.

- Date: 2024-07-19
- Lab: ElevenLabs
- Type: Speech
- Weights: Proprietary
- Modalities: text → audio
- Available at launch: API, Consumer app
- Permalink: https://theaiengineering.com/releases/2024/07#eleven-turbo-v2-5

Highlights:

- First support for Vietnamese, Hungarian and Norwegian
- Three times faster in non-English languages
- Later outclassed by Flash v2.5

Source: [Introducing Turbo v2.5](https://elevenlabs.io/blog/introducing-turbo-v25)

### GPT-4o mini

GPT-4o mini is OpenAI's cost-efficient small model, supporting text and vision in the API and replacing GPT-3.5 in ChatGPT for Free, Plus, and Team users. It is the first API model to apply OpenAI's instruction hierarchy method.

- Date: 2024-07-18
- Lab: OpenAI
- Type: Small model
- Weights: Proprietary
- Modalities: text, image → text
- Context window: 128,000 tokens
- Available at launch: API, Consumer app
- Permalink: https://theaiengineering.com/releases/2024/07#gpt-4o-mini

Highlights:

- API model ID gpt-4o-mini-2024-07-18
- 128K context window, up to 16K output tokens, knowledge up to October 2023
- Available in the Assistants, Chat Completions, and Batch APIs; fine-tuning launched Jul 23, 2024
- ChatGPT Enterprise access began the following week

Source: [GPT-4o mini: advancing cost-efficient intelligence](https://openai.com/index/gpt-4o-mini-advancing-cost-efficient-intelligence/)

### Granite Code 128K (3B, 8B)

Long-context versions of the 3B and 8B Granite Code models that extend the context window to 128K tokens, released as base and instruct checkpoints.

- Date: 2024-07-18
- Lab: IBM
- Type: Coding
- Weights: Open weights
- Modalities: text → text
- Context window: 128,000 tokens
- Parameters: 3B and 8B
- License: Apache-2.0
- Available at launch: Open weights
- Permalink: https://theaiengineering.com/releases/2024/07#granite-code-128k

Highlights:

- Extends Granite Code context to 128K tokens
- Base and instruct checkpoints for each size
- Released under Apache 2.0

Source: [ibm-granite/granite-8b-code-instruct-128k · Hugging Face](https://huggingface.co/ibm-granite/granite-8b-code-instruct-128k)

### Mistral NeMo

12B model built with NVIDIA that offers a 128K-token context window, multilingual skills and a drop-in replacement path for systems using Mistral 7B.

- Date: 2024-07-18
- Lab: Mistral AI
- Type: LLM
- Weights: Open weights
- Modalities: text → text
- Context window: 128,000 tokens
- Parameters: 12B
- License: Apache-2.0
- Available at launch: Open weights, API
- Permalink: https://theaiengineering.com/releases/2024/07#mistral-nemo

Highlights:

- Introduced the Tekken tokenizer, trained on more than 100 languages
- Trained with quantization awareness for FP8 inference
- Base and instruct checkpoints under Apache 2.0

Source: [Mistral NeMo | Mistral AI](https://mistral.ai/news/mistral-nemo)

### Codestral Mamba

Mamba-architecture code model designed with help from Albert Gu and Tri Dao, offering linear-time inference and tested on in-context retrieval up to 256K tokens.

- Date: 2024-07-16
- Lab: Mistral AI
- Type: Coding
- Weights: Open weights
- Modalities: text → text
- Context window: 256,000 tokens
- Parameters: 7.3B
- License: Apache-2.0
- Available at launch: Open weights, API
- Permalink: https://theaiengineering.com/releases/2024/07#codestral-mamba

Highlights:

- Instructed model with 7,285,403,648 parameters
- Released under Apache 2.0
- API model ID codestral-mamba-2407

Source: [Codestral Mamba | Mistral AI](https://mistral.ai/news/codestral-mamba)

### Mathstral 7B

STEM-specialized model built on Mistral 7B for multi-step mathematical reasoning, produced in collaboration with Project Numina and released with open weights.

- Date: 2024-07-16
- Lab: Mistral AI
- Type: LLM
- Weights: Open weights
- Modalities: text → text
- Context window: 32,000 tokens
- Parameters: 7B
- License: Apache-2.0
- Available at launch: Open weights
- Permalink: https://theaiengineering.com/releases/2024/07#mathstral

Highlights:

- Instructed model for math and STEM subjects
- Weights hosted on Hugging Face
- Can be fine-tuned with mistral-finetune

Source: [MathΣtral | Mistral AI](https://mistral.ai/news/mathstral)

### SmolLM

Family of small language models trained on SmolLM-Corpus, a curated dataset that combines synthetic textbooks, educational Python code and educational web data.

- Date: 2024-07-16
- Lab: Hugging Face
- Type: Small model
- Weights: Open weights
- Modalities: text → text
- Parameters: 135M, 360M and 1.7B
- License: Apache-2.0
- Available at launch: Open weights
- Permalink: https://theaiengineering.com/releases/2024/07#smollm

Highlights:

- Released in 135M, 360M and 1.7B sizes
- Training corpus includes Cosmopedia v2 and Python-Edu
- Released under Apache 2.0

Source: [SmolLM - blazingly fast and remarkably powerful](https://huggingface.co/blog/smollm)

### Granite 20B FunctionCalling

Function calling model fine-tuned from Granite 20B Code Instruct with a multi-task training approach to bring tool-calling abilities to the Granite family.

- Date: 2024-07-09
- Lab: IBM
- Type: LLM
- Weights: Open weights
- Modalities: text → text
- Parameters: 20B
- License: Apache-2.0
- Available at launch: Open weights
- Permalink: https://theaiengineering.com/releases/2024/07#granite-20b-functioncalling

Highlights:

- Based on granite-20b-code-instruct
- Trained with a multi-task function calling approach
- Released under Apache 2.0

Source: [ibm-granite/granite-20b-functioncalling · Hugging Face](https://huggingface.co/ibm-granite/granite-20b-functioncalling)

### Kolors

Kolors is Kuaishou's large text-to-image model trained on billions of text-image pairs, bilingual in Chinese and English, open-sourced at the 2024 World Artificial Intelligence Conference.

- Date: 2024-07-06
- Lab: Kuaishou
- Type: Image
- Weights: Open weights
- Modalities: text → image
- License: Kolors Model License Agreement
- Available at launch: Open weights
- Permalink: https://theaiengineering.com/releases/2024/07#kolors

Highlights:

- Bilingual Chinese and English prompts with a 256-token text context
- ControlNet, inpainting, and IP-Adapter variants released in July 2024
- Weights free for academic research; commercial use requires registration

Source: [GitHub - Kwai-Kolors/Kolors](https://github.com/Kwai-Kolors/Kolors)

### CodeGeeX4-ALL-9B

CodeGeeX4-ALL-9B is the open-weights version of Z.ai's CodeGeeX4 code model series, continually trained on GLM-4-9B. A single model supports code completion and generation, code interpreter, web search, function calling, and repository-level code Q&A.

- Date: 2024-07-05
- Lab: Z.ai
- Type: Coding
- Weights: Open weights
- Modalities: text → text
- Context window: 128,000 tokens
- Parameters: 9B
- License: CodeGeeX4 License
- Available at launch: Open weights
- Permalink: https://theaiengineering.com/releases/2024/07#codegeex4

Highlights:

- 128K context length
- Available on Hugging Face, ModelScope, WiseModel, and Ollama
- Weights under the CodeGeeX4 License

Source: [CodeGeeX4: Open Multilingual Code Generation Model (GitHub)](https://github.com/zai-org/CodeGeeX4)

## Cite and reuse this data

These release records are free to reuse under CC BY 4.0, including commercially. Credit The AI Engineering and link to this page's URL. Dates are the labs' official announcement dates, and every record links to its source. List fields in the CSV are separated by " | ".

The AI Engineering Editors. "AI models released in July 2024." The AI Engineering, updated Jul 31, 2024. https://theaiengineering.com/releases/2024/07

HTML link:

```html
<a href="https://theaiengineering.com/releases/2024/07">AI models released in July 2024</a> (The AI Engineering)
```

- Title: AI models released in July 2024
- Author: The AI Engineering Editors
- Publisher: The AI Engineering
- Updated: 2024-07-31
- URL: https://theaiengineering.com/releases/2024/07
- Markdown: https://theaiengineering.com/releases/2024/07.md
- License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)
- Full dataset as CSV: https://theaiengineering.com/releases.csv
- Full dataset as JSON: https://theaiengineering.com/releases.json
