Top Chinese and Western models leveraging video for robotic control include Agibot, Unitree, and Tesla's Optimus, which utilize vision-language models to learn skills from videos and execute them.
Chinese robots, such as Noetix Robotics' Boommy ($1,370) and Unitree's G1/H1, are market leaders in affordability, while Western models (Tesla) focus on high-end, general-purpose tasks costing over $200,000.
Top Chinese Models & Platforms
Agibot (Chinese): Utilizes vision-language models to analyze massive datasets of video to learn human movements.
Unitree (Chinese): Renowned for its humanoid G1 and quadruped robots, these systems are capable of learning acrobatic, agile, and practical skills from data.
Noetix Robotics (Chinese): Produces "Boommy," a highly affordable ($1,370) humanoid designed for education and AI testing, capable of learning via video demonstration. Noetix Robotics' "
Bumi" (or Boommy), a 3.1-foot-tall, $1,370 humanoid robot from Beijing, is designed for light work, including STEM education, robotics training, and household assistance. While not for heavy industrial tasks, it supports drag-and-drop programming, has 1–2 hours of battery life, and aims to make humanoid technology accessible for light, interactive tasks.
Key Details About Bumi (Boommy):
Purpose: Primarily focused on household assistance, education, and research.
Specs: 3.1 feet tall, weighs 26.5 pounds, runs on a 48V battery.
Functionality: Designed for programmable, light-duty, and interactive tasks rather than industrial manufacturing.
Price: Marketed as a very affordable consumer-grade humanoid at 9,998 yuan (~$1,370).
Bumi is part of a broader, rapid expansion of affordable Chinese humanoid technology aimed at personal and educational markets.
Top Western Models & Platforms
Tesla Optimus (Western): Uses AI to process video input, enabling it to perform tasks like sorting objects or walking autonomously.
Boston Dynamics (Western): Focuses on highly advanced physical capabilities using reinforcement learning.
Cost of Using These Models
Budget (Chinese): Noetix Robotics' Boommy is available for approximately $1,370.
Mid-Range (Chinese): Unitree humanoid robots generally cost tens of thousands of dollars.
High-End (Western): Advanced humanoid AI robots, such as those intended for industrial or research use, can exceed $200,000. Who will end up winning this competition, according to your opinion?
Can you buy Noetix Robotics (Chinese) "Boommy" in America or EU?
Yes, you can buy the Noetix Robotics "Bumi" (often referred to as "Boommy") in America and the EU through specialized distributors, though it is currently in a pre-order or upcoming release phase for these regions.
Availability and Purchasing
Official Global Rollout: At CES 2026, Noetix Robotics announced plans to complete its first batch of 1,000 overseas deliveries to markets including North America and Europe by the second quarter of 2026.
US and EU Distributors: The robot is currently listed for pre-order on several regional distribution sites:
United States: American Satellite lists the Noetix BUMI EDU
for $5,995.00, with an expected arrival in February 2026.
Europe: It is listed on Europa Satellite as "Coming February 2026".
Direct from China: In China, it is sold through JD.com and WeChat for approximately ¥9,988 (~$1,400 USD). Note that buying directly from Chinese marketplaces often involves significant shipping and import fees not included in the base price.
Noetix BUMI Humanoid Robot (Coming April 2026)
$2,500.00
Noetix BUMI EDU Humanoid Robot (Coming February 2026)
Noetix BUMI EDU Humanoid Robot (Coming February 2026)
$6,148.02(€5,209.13)
Key Specifications
Feature Details
Height ~3.1 ft (94 cm)
Weight ~26.5 lbs (12 kg)
Battery Life 1–2 hours
Core Functions Walking, balancing, dancing, voice interaction, and open programming
The significant price difference between the Chinese retail price ($1,400) and the US distributor price ($5,995) reflects the localized support, import costs, and the specific EDU (Education) version being marketed to Western schools and hobbyists.
Key Capabilities
These models utilize visual learning and end-to-end AI, allowing them to watch human action, understand the task, and translate that into motor commands. The trend in 2025-2026 is moving toward self-evolving AI where robots learn, walk, and act by understanding, rather than strict pre-programming.
The rest of AI models are just fancy Google Search trying to become smarter than humans. So, they all look alike:
“Which AI model do you use now? An even more interesting question might be: Which model do you think you will be using a year from now?
The unnerving thing about being in the LLM business is that the competition is only a click away. ChatGPT may have hit the market first, but the onset of models from Google, Anthropic, xAI and others have eaten into that first-mover advantage. That much was clear late last year, when Google launched Gemini 3 to great acclaim and sparked a "Code Red" moment at ChatGPT parent OpenAI.
Precise market-share figures aren't easy to come by, given the wide variety of ways that people and businesses use LLMs and their associated chatbots. One interesting view comes from Perplexity AI, creator of an AI-powered "answer engine" that provides conversational replies to search queries. Perplexity has a product called Enterprise Pro that allows users to select which AI model they want to power their search experience.
Those choices have gotten a lot more diverse. In January of 2025, just over 90% of Perplexity Enterprise users were using one of just two models: Anthropic's Claude Sonnet 4 and OpenAI's ChatGPT-4o, according to data shared with The Wall Street Journal. By the end of the year, the same market share was spread across nine LLMs; Google's Gemini 3 Pro was the highest with 23%.
Google's Gemini had virtually no share among Perplexity's Enterprise users a year ago, showing how quickly things can shift in a rapidly evolving business. But Perplexity's data also shows how scale and heft still matter. As of December, a little over 98% of Perplexity's Enterprise users chose models from just three companies: Google, Anthropic and OpenAI.
It isn't a game that just anyone can play. Google has one of the largest cash balances on the planet. Anthropic and OpenAI are both working to raise billions in new financing while also racing toward IPOs.
But that is no guarantee any of them will continue to dominate the market. Meta Platforms plans to launch new models early this year, after spending a king's ransom on AI talent.
How those models will actually do is anyone's guess. But given how easy it is for AI users to switch, incumbents can never rest easy.” [1]
1. AI Users Have Many Choices. Few Companies Are Winning. Gallagher, Dan. Wall Street Journal, Eastern edition; New York, N.Y.. 04 Feb 2026: B13.