xAI releases Grok 4.1 and Grok 4.1 Fast
On 17 November 2025 xAI announced Grok 4.1, after a silent rollout from 1 to 14 November. Two days later it released Grok 4.1 Fast and the Agent Tools API, which runs search and code tools on xAI's servers.
Total 671 · Clear filters
Baidu unveils the omni-modal ERNIE 5.0 at Baidu World 2025
On 13 November 2025 Baidu unveiled ERNIE 5.0 at Baidu World 2025, describing it as a natively omni-modal model. It ranked eighth on the LMArena text leaderboard after leaving preview.
OpenAI releases GPT-5.1 with personality presets
On 12 November 2025 OpenAI released GPT-5.1 Instant and GPT-5.1 Thinking in ChatGPT, giving the chatbot a warmer default tone and eight selectable personality presets.
iFlytek releases Spark X1.5
iFlytek released Spark X1.5 on 6 November 2025 at its 1024 Developer Festival. The MoE model has 293 billion total parameters, doubled inference efficiency over X1, and claimed 95% of GPT-5's overall level.
XPENG unveils VLA 2.0, a Robotaxi plan and a new IRON robot at AI Day
At its 2025 AI Day in Guangzhou, XPENG presented VLA 2.0, a Robotaxi, a next-generation IRON humanoid and two ARIDGE flying systems, and said it would open-source VLA 2.0 with Volkswagen as first customer.
XPENG unveils new-generation IRON
At its 2025 AI Day in Guangzhou, XPENG showed a new-generation IRON humanoid robot with 82 degrees of freedom, three Turing AI chips and an all-solid-state battery, and said it targets volume production by the end of 2026.
Moonshot AI releases Kimi K2 Thinking
Moonshot AI released Kimi K2 Thinking in November 2025, a reasoning model built on the trillion-parameter Kimi K2 layout with INT4 quantization for its mixture-of-experts parts and long tool-use chains.
Baidu open-sources ERNIE-4.5-VL-28B-A3B-Thinking
In November 2025 Baidu released ERNIE-4.5-VL-28B-A3B-Thinking, an open-weight vision-language model that reasons with images — zooming into details and calling image search mid-thought — and reports a ChartQA score above Gemini 2.5 Pro and GPT-5.
Cursor 2.0 and Composer
On 29 October 2025, Cursor released Cursor 2.0 and Composer, its first coding model, with a new agent-centered interface, parallel agents, and a native browser tool for testing changes.
NVIDIA releases Isaac GR00T N1.6 with the Cosmos open models
On 28 October 2025 NVIDIA released Isaac GR00T N1.6, an open model for humanoid robots that adds a reasoning vision-language model, Cosmos Reason, so vague instructions can be turned into step-by-step action plans.
1X opens pre-orders for the NEO home robot
On 28 October 2025, 1X announced NEO, which it calls the first consumer-ready humanoid robot for the home, priced at US$20,000 for Early Access with 2026 delivery in the US and a US$499 monthly subscription later.
LangGraph 1.0 released
LangGraph 1.0, generally available on 23 October 2025, was called by LangChain the first stable major release for durable agent frameworks, adding resumable state, built-in persistence and human-in-the-loop APIs, with no breaking changes promised until 2.0.
OpenAI launches the ChatGPT Atlas browser
On 21 October 2025 OpenAI released ChatGPT Atlas, a macOS browser built around ChatGPT, with agent mode and browser memories. It arrived as a challenge to Google Chrome and was shut down on 9 August 2026.
Unitree releases the H2 bionic humanoid
On 20 October 2025 Unitree announced the H2, a bionic humanoid robot with a face, 31 joints, a 180 cm frame and a 70 kg body, shown in a video performing ballet, punches, side kicks and a runway walk.
DeepSeek releases DeepSeek-OCR, compressing text into images
On 20 October 2025 DeepSeek released DeepSeek-OCR under the MIT License. The model showed that rendering text as images can reduce the number of vision tokens needed for OCR, with decoding precision of about 97 percent at compression under 10x.
Anthropic launches Claude Code on the web
On 20 October 2025 Anthropic launched Claude Code on the web, a research preview that lets developers hand coding tasks from a browser to Claude, running on Anthropic-managed cloud infrastructure.
Anthropic introduces Agent Skills
On 16 October 2025 Anthropic introduced Agent Skills, folders of instructions, scripts and resources that Claude loads when a task calls for them. On 18 December 2025 the format was published as an open standard.
Anthropic releases Claude Haiku 4.5
On 15 October 2025 Anthropic released Claude Haiku 4.5 to all users. It cost $1 and $5 per million tokens and matched Sonnet 4's coding performance at one-third the cost.
Microsoft AI introduces MAI-Image-1
On 13 October 2025 Microsoft AI announced MAI-Image-1, its first image model built entirely in-house. It reached LMArena's top 10 and began rolling out in Bing Image Creator and Copilot on 4 November 2025.
Figure introduces Figure 03
On 9 October 2025 Figure introduced Figure 03, its third-generation humanoid robot, built around a redesigned body, a new manufacturing process at its BotQ facility, and much lower unit cost.
AI21 Labs releases Jamba Reasoning 3B for on-device use
On 8 October 2025 AI21 Labs released Jamba Reasoning 3B, a 3-billion-parameter open reasoning model under Apache 2.0 that it aimed at phones, laptops and PCs.