Chinese companies continue to release AI models that rival the capabilities of systems developed by OpenAI and other US-based AI companies.
this week, Mini Maxa startup backed by Alibaba and Tencent He grew up About $850 million in venture capital and valued at more than $2.5 billion, For the first time three New models: MiniMax-Text-01, MiniMax-VL-01, T2A-01-HD. MiniMax-Text-01 is a text-only model, while MiniMax-VL-01 can understand both images and text. Meanwhile, the T2A-01-HD generates sound, specifically speech.
MiniMax claims that MiniMax-Text-01, with its 456 billion parameters, performs better than models like Google’s recently unveiled Gemini 2.0 Flash in benchmarks like MATH and SimpleQA, which measure a model’s ability to answer math and factual problems. Existing questions. The parameters roughly correspond to the model’s problem-solving skills, and models with more parameters generally perform better than those with fewer.
As for the MiniMax-VL-01, MiniMax says it competes with Anthropic’s Claude 3.5 Sonnet in assessments that require multimodal understanding, such as ChartQA, which tasks models with answering queries about graphs and charts (e.g., “What is the maximum value of the model?” ? The orange line in this graph?”). The MiniMax-VL-01 certainly doesn’t outperform the Gemini 2.0 Flash on many of these tests. OpenAI’s GPT-4o and Meta’s Llama 3.1 beat it in many games as well.
It should be noted that MiniMax-Text-01 has a very large context window. Model context, or context window, refers to the inputs (for example, text) that the model takes into account before generating the output (additional text). With a contextual window containing 4 million symbols, the MiniMax-Text-01 can parse about 3 million words at once — or just over five copies of “War and Peace.”
For context (no pun intended), the MiniMax-Text-01’s context window is about 31 times the size of GPT-4o and Llama 3.1.
The latest MiniMax model released this week, the T2A-01-HD, is a speech-enhanced audio generator. The T2A-01-HD can create synthetic voice with adjustable tempo, pitch, and tenor in about 17 different languages, including English and Chinese, and reproduce audio from just 10 seconds of audio recording.
MiniMax has not published benchmark results comparing the T2A-01-HD to other sound generating models. But to this reporter’s ear, the T2A-01-HD’s output sounds on par with audio models from dead And startups like PlayAI.
With the exception of T2A-01-HD, which is available exclusively through the MiniMax API and Hailuo AI platform, new MiniMax models can be downloaded from GitHub and the Hugging Face AI development platform.
However, just because models are available “openly” does not mean that they are not closed in certain aspects. MiniMax-Text-01 and MiniMax-VL-01 are not real source, meaning MiniMax has not released the components (e.g., training data) needed to recreate them from scratch. Furthermore, it is subject to a restricted MiniMax license, which prohibits developers from using models to improve competing AI models and requires that platforms with more than 100 million monthly active users request a special MiniMax license.
MiniMax was founded in 2021 by former employees of SenseTime, one of the largest AI companies in China. The company’s projects include apps like Talkie, an AI-powered role-playing platform similar to Character AI, and text-to-video models released by Hailuo’s MiniMax.
Some MiniMax products have become the subject of minor controversy.
Talkie, which was pulled from Apple’s App Store in December for unspecified “technical” reasons, features AI-powered avatars of public figures, including Donald Trump, Taylor Swift, Elon Musk, and LeBron James, none of whom appear to have approved of Appear in the application. Application.
In December, Broadcast Magazine I mentioned That MiniMax’s video generators can reproduce the logos of British TV channels suggests that MiniMax’s models were trained on the content of those channels. It is said that MiniMax He is being sued By iQiyi, a Chinese video streaming service that claims the MiniMax was illegally trained on iQiyi’s copyrighted recordings.
The new MiniMax models arrive days after the outgoing Biden administration proposed tougher export rules and restrictions on artificial intelligence technologies for Chinese projects. Companies in China have already been banned from purchasing advanced AI chips, but if the new rules take effect as written, companies will face tougher restrictions on both the semiconductor technology and models needed to power advanced AI systems.
On Wednesday, the Biden administration Announce Additional measures focused on keeping advanced chips out of China. Chip foundries and packaging companies that want to export certain chips will be subject to broader licensing requirements unless they exercise greater scrutiny and due diligence to prevent their products from reaching Chinese customers.