Total 1,098
On 1 September 2025 Meituan's LongCat team released LongCat-Flash, a model with 560 billion total parameters that activates 18.6 to 31.3 billion per token, ships open weights under the MIT license and is aimed at agentic use.
Four Chinese departments issued measures requiring labels on AI-generated text, audio, images and video, in explicit and file-metadata form, effective 1 September 2025.
In September 2025, Li Auto began rolling out its VLA Driver model via OTA 8.0 to all AD Max users, reporting sharp increases in assisted-driving usage and take-over intervals.
In September 2025 Disney, Universal and Warner Bros. Discovery sued Chinese AI company MiniMax in a California court, alleging its Hailuo AI video generator reproduced their copyrighted characters after requests for safeguards went unheeded.
In late September 2025 Zhipu released GLM-4.6 under the MIT License, with a 200K-token context window, stronger coding and agent abilities, and support for Chinese domestic chips.
In September 2025 Alibaba open-sourced Qwen3-Omni, an omni-modal model under Apache 2.0 that processes text, audio and video and claimed state-of-the-art results on many audio benchmarks.
In September 2025 Alibaba released Qwen3-VL, a vision-language model family under Apache 2.0 whose flagship handles hours-long video, operates PC and mobile interfaces and does 3D grounding.
In September 2025 Alibaba's Qwen team released Qwen3-Next, an 80B-parameter open model with 3B parameters active per token, built on a hybrid of linear attention and sparse mixture-of-experts layers.
In September 2025 Alibaba released Qwen3-Max, a proprietary model with over 1 trillion parameters trained on about 36 trillion tokens. It became the flagship of Alibaba's closed Max tier, released alongside several open Qwen3 models.
Kuaishou released Kling 2.5 Turbo in late September 2025; ten days later it ranked No. 1 in both text-to-video and image-to-video on Artificial Analysis, per-video costs fell nearly 30%, and Kling AI revenue exceeded RMB 300 million in Q3 2025.
ByteDance Seed released Seedream 4.0 in September 2025, with text-to-image generation, image editing and multi-image composition in one model, producing images from 1K to 4K.
A class action by authors against Anthropic ended in a $1.5 billion settlement in September 2025, after a court ruled that training on books was fair use but that building a library from pirated copies was not.
On 28 August 2025 xAI released Grok Code Fast 1, a fast reasoning model for agentic coding. It launched free for a limited time inside several popular coding tools.
On 28 August 2025 Microsoft AI announced its first in-house models, the MAI-Voice-1 speech model and the MAI-1-preview language model. Coverage read the release as a first step to reduce dependence on OpenAI.
On 26 August 2025 Google added a new image editing model, nicknamed Nano Banana, to the Gemini app and called it the top-rated image editing model on the LMArena image-edit leaderboard. Developers could use the same model as Gemini 2.5 Flash Image.
On 25 August 2025 Anthropic began a limited pilot of Claude in Chrome, a browser extension that can act on web pages, publishing prompt-injection test results alongside a rollout that widened to more plans by December.
On 22 August 2025, Meta said it would license Midjourney's AI image and video generation technology, with Meta research teams collaborating with the startup. Financial terms were not disclosed, and Midjourney said it remains independent.
On 21 August 2025 DeepSeek released DeepSeek-V3.1, its first hybrid model, which runs in either a thinking or a non-thinking mode under the MIT License.
On 21 August 2025 Cohere released Command A Reasoning, a hybrid reasoning model for complex agentic tasks. It has 111 billion parameters, a 256k-token context and users can switch its thinking on or off.
On 20 August 2025 ByteDance Seed released the Seed-OSS-36B models on Hugging Face under Apache-2.0. The 36B models were trained on 12T tokens and support a native context of up to 512K tokens.
The 2025 World Humanoid Robot Games opened at Beijing's Ice Ribbon on 14 August 2025, bringing hundreds of humanoid robots and teams from more than a dozen countries into competition across athletic, performance and workplace-style scenario events.
