Total 671 · Clear filters
On 13 April 2023 AWS announced Amazon Bedrock, a service giving API access to foundation models from several providers, alongside its first Titan models. Bedrock became generally available on 28 September 2023.
On 10 April 2023 SenseTime unveiled its SenseNova foundation model set and the SenseChat chatbot, with a live demonstration by CEO Xu Li.
On 5 April 2023 Meta released the Segment Anything Model (SAM), a promptable image segmentation model, together with the SA-1B dataset of over 1.1 billion masks. Meta presented it as a step toward a foundation model for image segmentation.
In April 2023 Alibaba launched Tongyi Qianwen as a beta, and in September 2023 it opened the model to the public. The project later took the international name Qwen.
On 16 March 2023 Robin Li unveiled ERNIE Bot, Baidu's chatbot, to invited testers at a Beijing event. Baidu's Hong Kong shares fell as much as 10% that day, and closed down 6.4%.
On 16 March 2023 Microsoft introduced Microsoft 365 Copilot, pairing GPT-4 and other large language models with business data, and made it generally available in November.
Midjourney released version 5 of its text-to-image service in alpha on 15 March 2023, after founder David Holz said it would arrive that week. Test images showed a higher default resolution and a wider range of styles, and point releases followed through the summer.
On 14 March 2023 Anthropic made its AI assistant Claude available more widely, after a closed alpha, in two versions. The system was built on research into helpful, honest and harmless AI, and launch partners included Quora and Notion.
On 14 March 2023 OpenAI announced GPT-4, a multimodal model that accepts image and text input and produces text output. OpenAI reported human-level results on several professional and academic benchmarks but disclosed little about the model's design.
On 9 March 2023 AI21 Labs announced Jurassic-2 and a set of Task-Specific APIs. The models came in three sizes and were free to use until 1 May 2023.
On 2 March 2023, Figure came out of stealth and introduced Figure 01, a planned bipedal humanoid aimed at manufacturing, logistics, warehousing and retail work; the images released were renderings, not working hardware.
On 1 March 2023 OpenAI released gpt-3.5-turbo, the model behind ChatGPT, as an API priced at $0.002 per 1,000 tokens, and opened a Whisper speech API the same day.
In March 2023 Runway introduced Gen-2, a model that turns text, images or video clips into video, released publicly that June. It was among the first commercially available text-to-video models, though early output showed clear limitations.
On 24 February 2023 Meta announced LLaMA, a family of foundation language models with 7 to 65 billion parameters. Meta released it to approved researchers under a noncommercial license, and the paper described training on public data only.
On 7 February 2023 Microsoft unveiled an AI-powered Bing and Edge, built on a new OpenAI model, whose chat feature was launched as a waitlisted preview called Bing Chat.
On 6 February 2023 Google announced Bard, an experimental conversational AI service built on LaMDA. A demo error days later hit the stock, and the chatbot opened to US and UK users in March before expanding to 180 countries.
In February 2023 Runway released Gen-1, a video-to-video model that restyles existing footage from text or image prompts while keeping the source structure, an early step from image generation toward generative video editing.
On 1 February 2023 OpenAI launched ChatGPT Plus, a $20-a-month pilot subscription offering access at peak times, faster responses and priority access to new features, while free access was kept.
In 2023 Sanctuary AI unveiled Phoenix, its sixth-generation general-purpose humanoid robot, less than two months after announcing its first customer deployments, with total funding above C$100 million.
ChatGLM-6B is a 6.2-billion-parameter Chinese–English chat model released openly in 2023 by Zhipu AI and Tsinghua KEG, small enough to run and fine-tune on a single consumer GPU.
In December 2022, Google researchers posted RT-1, a Robotics Transformer trained on real robot data; they reported over 700 instructions at a 97% success rate and released the code.
