AI models released in August 2023
12 releases from 9 labs in August 2023, including Code Llama, babbage-002 and davinci-002 and Eleven Multilingual v2. Every entry is dated to the lab's official announcement and links to it.
- Releases
- 12
- Labs
- 9
- Open weights
- 7
August 2023
12 releasesCode Llama
Code Llama is a family of code-specialized Llama 2 models, released in base, Python and Instruct variants for code generation and completion.
- Modalities
- text → text
- Context
- 100K context
- Parameters
- 7B, 13B, 34B
- License
- Llama 2 Community License
- At launch
- Open weights
- Stable generations with up to 100,000 tokens of context
- Code Llama - Python was further fine-tuned on 100B tokens of Python code
- Released under the same community license as Llama 2
babbage-002 and davinci-002
babbage-002 and davinci-002 are GPT base models without instruction tuning, released as replacements for the original GPT-3 base models in the legacy Completions API and usable as base or fine-tuned models.
- Modalities
- text → text
- At launch
- API
- Replace ada, babbage, curie, and davinci, which were set to be turned off on Jan 4, 2024
- Announced together with fine-tuning for GPT-3.5 Turbo
- 16,384 max output tokens and Sep 1, 2021 knowledge cutoff per the docs model pages
Eleven Multilingual v2
Foundational multilingual speech synthesis model supporting 29 languages, released as ElevenLabs officially ended its beta phase and positioned as its lifelike model with rich emotional expression.
- Modalities
- text → audio
- At launch
- Consumer app, API
- Supports 29 languages
- Model ID eleven_multilingual_v2
- Remains a flagship text-to-speech model in the API
IDEFICS
Open-access visual language model from Hugging Face that reproduces DeepMind's closed Flamingo model, accepting arbitrary sequences of images and text and generating text responses.
- Modalities
- text, image → text
- Parameters
- 9B and 80B
- At launch
- Open weights
- Trained partly on the new OBELICS dataset of 141 million interleaved documents
- Built on CLIP-ViT-H-14 and LLaMA 65B
- Released in 9B and 80B variants with instruct versions
Ideogram 0.1
Ideogram's first publicly launched text-to-image model, introduced with the announcement of Ideogram AI and opened to everyone for free one week later.
- Modalities
- text → image
- At launch
- Consumer app
- Launched with the announcement of Ideogram AI
- Opened to everyone for free on August 29, 2023
Qwen-VL
Qwen-VL is Qwen's first open vision-language model, released with Qwen-VL-Chat. It uses 448 by 448 image resolution for fine-grained text recognition, document question answering, and bounding box annotation.
- Modalities
- text, image → text
- License
- Tongyi Qianwen License Agreement
- At launch
- Open weights
- 448 by 448 resolution, higher than the 224 by 224 used by other open models at the time
- Released on ModelScope and Hugging Face with a paper
- Commercial use allowed under its license
SeamlessM4T
SeamlessM4T is a single multimodal model for speech and text translation and speech recognition, covering nearly 100 input languages and 35 speech output languages plus English.
- Modalities
- text, audio → text, audio
- License
- CC-BY-NC-4.0
- At launch
- Open weights
- Speech recognition for nearly 100 languages
- Speech-to-speech translation into 35 languages plus English
- Released publicly under CC BY-NC 4.0
Claude Instant 1.2
Claude Instant 1.2 is an updated version of Anthropic's faster, lower-priced Claude Instant model that incorporates strengths of Claude 2, with gains in math, coding, reasoning, and safety and longer, more structured responses.
- Modalities
- text → text
- At launch
- API
- API model ID claude-instant-1.2
- Available through the Anthropic API
- Improves on Claude Instant 1.1 in quote extraction, multilingual capabilities, and question answering
StableCode (Alpha)
Stability AI's first code models: a base model trained on BigCode's The Stack, an instruction-tuned model and a long-context completion model aimed at single and multi-line autocomplete.
- Modalities
- text → text
- Parameters
- 3B
- License
- Apache-2.0
- At launch
- Open weights
- Three models: base, instruction-tuned and long-context completion
- Completion model released under Apache 2.0
- 3B parameters
Skylark
Skylark was ByteDance's first-generation language model family on Volcano Engine's Ark platform, released in pro and lite tiers alongside Skylark-plus and later joined by Skylark-chat, the same model used in the Doubao app.
- Modalities
- text → text
- At launch
- API
- Skylark-pro and Skylark-lite V1.0 listed as new releases
- Skylark-chat, matching the Doubao app, added September 26, 2023
- Version 2.0 in October 2023 gave Skylark-lite 8K and Skylark-pro 4K context
- Retired on August 20, 2024
Qwen-7B
Qwen-7B is the first open model from Alibaba's Qwen (Tongyi Qianwen) team, released with the Qwen-7B-Chat assistant model on ModelScope and Hugging Face.
- Modalities
- text → text
- Parameters
- 7B
- License
- Tongyi Qianwen License Agreement
- At launch
- Open weights
- Pre-trained on 2.4T tokens
- Released with a technical memo on training details
- Weights under the Tongyi Qianwen License Agreement
AudioCraft (MusicGen and AudioGen)
Meta open-sourced AudioCraft, its generative audio framework, releasing model weights and code for MusicGen text-to-music, AudioGen text-to-sound and an improved EnCodec decoder.
- Modalities
- text → audio
- License
- CC-BY-NC-4.0
- At launch
- Open weights
- Model weights released under CC-BY-NC 4.0 for research use
- Training code released under the MIT license
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 " | ".
Citation
The AI Engineering Editors. "AI models released in August 2023." The AI Engineering, updated Aug 24, 2023. https://theaiengineering.com/releases/2023/08
HTML link
<a href="https://theaiengineering.com/releases/2023/08">AI models released in August 2023</a> (The AI Engineering)- Full dataset
- 1,167 releases from 55 labs
- Coverage
- to
- Updated
- License
- CC BY 4.0
- Publisher
- The AI Engineering