Chinese AI giant SenseTime has launched a new series of AI models, SenseNova 5.0, according to a new report published on April 23.
Among the news are a range of AI models for various purposes, including models for mobile phones and other devices, as well as models specialized in areas such as finance, programming and healthcare.
READ MORE: ChatGPT makes the EU election compass – that party AI had voted for
Better than GPT-4?
SenseTime demonstrated the capabilities of its AI models by comparing them to GPT-4 in tasks such as creative writing, logical reasoning and image analysis.
According to the report, one of the models, SenseChat V5, outperforms GPT-4 in almost all metrics.
SenseChat V5 can take in and remember 200,000 tokens, equivalent to around 140,000 words, at the same time and, according to the report, has made great strides in math and coding, among other things, compared to previous versions.
READ MORE: How to use the “ChatGPT killer” Claude 3 in Sweden – completely free
Text-to-video
The company also launched a new text-to-video tool, where users can create video sequences by describing them in text.
READ MORE: How to get better responses from ChatGPT – 5 unique tricks
READ MORE: How to recognize fake AI images – the signs many miss
The information and text in the article was produced using the AI tool Gemini. It is then processed and fact-checked by News24.