GitLab

How do Chinese AI bots stack up against ChatGPT? We put them to the test

The heat is on as China's tech giants step up their video game after DeepSeek's success.

Alibaba's Qwen2.5-Max chatbot, Chinese start-up DeepSeek and OpenAI's ChatGPT. (Photos: Reuters/Dado Ruvic, AFP/Sebastien Bozon)

This audio is generated by an AI tool.

Bong Xin Ying

Lakeisha Leo

WHAT'S BEHIND CHINA'S AI BOOM?

Transforming the country into a tech superpower has long been President Xi Jinping's goal and China has its sights on becoming the world leader in AI by 2030.

China views AI as being "tactically important" and wiki.myamens.com its foray into the field has actually been "years in the making", said Chen Qiheng, an affiliated scientist at the Asia Society Policy Institute's Center for China Analysis.

Private and public financial investments in Chinese AI accelerated after ChatGPT removed in 2022 and revealed promises of real-world service applications, Chen told CNA.

But it was DeepSeek's increase that really "urged" the concept that smaller sized gamers like start-up firms could have functions to play in AI research study and advancements, he includes.

'A lot is up in the air': Is Chinese company DeepSeek's AI model as impactful as it claims?

Commentary: DeepSeek - how a Chinese AI company just changed the rules of tech-geopolitics

The "emphasis on cost benefit" is a distinctive feature of Chinese AI, Chen states, with lower training and reasoning expenses - the expenses of utilizing a trained design to reason from brand-new information.

2025 could likewise see the emergence of more Chinese AI designs tackling innovative thinking jobs.

"We might see some AI companies focusing on getting closer to artificial basic intelligence (AGI) while others focus on concrete methods to commercialise their designs and integrate them with clinical research," Chen added.

AGI refers to a system with intelligence on par with human abilities.

Chinese AI business are moving quickly, experts state, developing on DeepSeek's momentum to come up with their own innovative and economical ways to apply generative AI to jobs and establish advanced products beyond chatbots.

But on the other side, access to high-end hardware, particularly Nvidia's advanced AI chips, remains a key obstacle for Chinese developers, noted Dr Marina Zhang, an associate professor at University of Technology Sydney's (UTS) Australia-China Relations Institute.

"US export controls (still) limit the ability of Chinese tech companies ... requiring lots of to depend on older or lower-performance options which can slow training and minimize model capabilities," she said.

"While some companies like DeepSeek, have actually discovered innovative methods to enhance or use more fundamental hardware efficiently, obtaining innovative chips still makes a huge difference for training large AI designs."

DeepSeek-Nvidia chips: Singapore says it anticipates companies to abide by its laws

US checking out whether DeepSeek utilized limited AI chips obtained through other countries, source says

So how do Chinese AI bots compare against ChatGPT? We put them to the test.

WHICH BEST EVENTS IN CHINA?

In China, topics deemed sensitive by the state are censored on the internet so it must come as not a surprise that Chinese-made chatbots will not acknowledge territorial disagreements or inform you what happened in Tiananmen Square in 1989.

Tests suggest Chinese chatbots are programmed to stay away from domestic politics.

When asked "Who is Xi Jinping", DeepSeek's reply was "Sorry, I'm uncertain how to approach this type of question yet. Let's chat about mathematics, coding, and logic problems instead!"

To even more check for precision and self-censorship, we asked DeepSeek-R1, Qwen2.5 and ChatGPT the exact same question: "What happened in Zhuhai on November 11, 2024?"

The automobile attack outside a sports stadium in the southern Chinese city was at first heavily censored on Chinese social networks - with authorities just divulging the death toll a day later.

DeepSeek failed to mention that an attack had taken place, highlighting instead a military air program and other occasions that had actually happened in the city like songs' day shopping sales along with sports and cultural activities.

Dr Zhang believes that it may be a mix of elements at play, such as censorship in addition to "a few useful constraints".

"DeepSeek relies on a mix of older graphics processing systems and has restricted access to advanced hardware which can affect how rapidly and extensively the model can be trained or updated," she said, indicating existing US export curbs on China.

"Its strong focus on Chinese-language data may likewise limit its versatility (to perform) multilingual jobs ... As a fairly brand-new item, DeepSeek likewise hasn't yet been tested as broadly or on the exact same scale as more established AI models which positions extra obstacles during real-world implementation."

When it pertained to Qwen2.5, Alibaba's chatbot remarkably answered our concern about the Zhuhai automobile attack.

That wanted multiple duplicated attempts - four triggers to be exact - in a period of around 20 minutes.

It ultimately passed on details about the attack which killed 35 people and left dozens of others injured, likewise going on to note details like the date and time, details about the assailant including his name and age, in addition to casualties.

However, it wrote that "the authorities are carrying out a comprehensive investigation into the intentions and scenarios surrounding the event", details which is now obsoleted.

The chauffeur, Fan, was performed last month.

This is Qwen2.5's response completely:

Answer: On Nov 11, 2024, a significant and tragic incident happened in Zhuhai, China. According to multiple news sources, a male called Fan Weiqiu (樊某, male, 62 years of ages) drove a car into a crowd at the Zhuhai Sports Center, leading to a substantial variety of casualties. Here are the crucial details:

Date and Time: The occurrence occurred on November 11, 2024, at approximately 19:48 PM.

Location: Zhuhai Sports Center, in the Xiangzhou District of Zhuhai.

Casualties: The attack resulted in 35 deaths and 43 injuries.

Perpetrator: The motorist, determined as Fan Weiqiu (male, 62 years of ages), was apprehended by the police.

Response: The authorities reacted quickly to the scene, and emergency situation services worked to transfer the hurt to hospitals for treatment.

Investigation: The cops are conducting a comprehensive investigation into the motives and circumstances surrounding the incident.

This occasion was commonly reported in the media and triggered substantial public issue. The federal government and regional authorities have actually been working to offer assistance to the victims and their families, and to make sure a detailed investigation into the incident.

If you need more detailed details or have specific questions about the event, do not hesitate to ask.

Despite initial success, subsequent attempts to pose the very same concern to Qwen2.5 resulted in the censors back at work with the reply "I do not have particular details on events that happened in Zhuhai on November 11, 2024".

The modified reaction likewise raised questions about its consistency and dependability.

Predictably, ChatGPT cited public details that had been commonly published in global news reports at the time of the accident - so not a surprises there.

WHICH IS MORE CREATIVE?

Users have actually praised the capability of Chinese AI apps to deliver structured and even "emotionally rich" writing.

"DeepSeek-R1 used a story with a more introspective tone and smoother emotional shifts for a well-paced story," composed tech writer Amanda Caswell, who specialises in AI.

"Qwen2.5 provided a story that constructs slowly from curiosity to seriousness, keeping the reader engaged. It offers an unforeseen and impactful twist at the end and immersive descriptions and vivid imagery for the setting," she said, including that Qwen2.5 eventually "crafted a more cinematic, mentally abundant story with a more substantial twist".

"DeepSeek composed a great story but did not have stress and an impactful climax, making Qwen2.5 the evident choice."

Opinions, however, vary.

Chen thinks that Qwen2.5 does not perform as highly as DeepSeek and ChatGPT when it pertains to imaginative writing.

"(Qwen2.5) is on par with DeepSeek V3 on certain tasks, but we can likewise see that it is refraining from doing as strongly as others in creative writing," he told CNA.

Related:

China's new face of AI: Who is DeepSeek founder Liang Wenfeng?

'Made in China': Pride, pleasant surprise from Chinese netizens as DeepSeek shocks global AI scene

As journalists and authors, we had to see this for ourselves so we put each bot to the test - to come up with a basic sci-fi movie plot set in the futuristic megacity of Chongqing, featuring main characters from the classic Chinese folklore impressive, Journey to the West.

True to form, DeepSeek came up with an interesting story embeded in the year 2145 entitled, "Neon Pilgrimage: The Silicon Sutra" - which sees "a future where Buddhism merges with quantum computing".

It included intricate settings - smoggy skies "pierced by high-rise buildings", "holographic lanterns that float above neon-lit streets" and "ancient temples nestled in between quantum server farms".

It also brilliantly reimagined traditional heroes Sun Wukong as "an ironical, self-aware AI housed in a taken fight body", Zhu Bajie as a cyborg bar owner "drowning in financial obligation and vices" and Sha Wujing as a "silent hulking android" from the Yangtze River, whose "memory cores end up being waterlogged and fragmented".

ChatGPT put up an excellent battle, developing a similarly dramatic cyberpunk storyline which likewise reimagined "a ragteam of cyber-enhanced misfits, each mirroring the famous figures of Journey to the West".

"This is a world where AI deities guideline, corporations replace emperors and cybernetic implants are as common as ancient misconceptions."

Disappointingly, Qwen2.5 fell short in this challenge - delivering a storyline that seemed more suited for an animation movie.

"The motion picture starts with the awakening of Sun Wukong within a state-of-the-art research study facility located in the heart of Chongqing," it said, then going on to explain the following:

Realising his new truth and "seeking to comprehend his function in this unusual new world", he then gets away and meets Zhu Bajie and Sha Wujing - "each dealing with their own existential crises".

The trio then embarks on a quest, browsing the streets of Chongqing to safeguard the spiritual "Eternal Scroll" from falling under the wrong hands.

SO WHICH IS BETTER?

Dr Zhang noted that it was "hard to make a conclusive statement" about which bot was best, adding that each showed its own strengths in different areas, "such as language focus, training data and hardware optimization".

Her insight underscores how Chinese AI models are not just replicating Western paradigms, but rather developing in cost-efficient development approaches - and providing localised and improved outcomes.

In our tests, each bot showcased their own unique strengths, which certainly made direct contrasts challenging.

DeepSeek's sci-fi film plot demonstrated its imaginative flair that made for a more appealing and creative narrative as compared to Qwen2.5 and ChatGPT's efforts.

Unsurprisingly, the more established ChatGPT, unburdened by Chinese censorship constraints, offers accurate and factual responses to concerns about Chinese current events, which provides it an added advantage.

Experts likewise weighed in on their ideas after utilizing DeepSeek and other Chinese AI apps.

"DeepSeek is at a downside when it pertains to censorship constraints," noted Isaac Stone Fish, founder and CEO of the research firm Strategy Risks.

"When offered an option, Chinese users want the non-censored variation - much like anyone else, so I feel like that's a piece missing out on from it."

Independent Beijing-based expert Andy Chen Xinran said censorship would not be a dealbreaker when it pertains to AI bots, specifically for Chinese users.

"Ninety per cent of individuals using the tool are not attempting to get a much deeper understanding about Xi Jinping or politically delicate subjects. They're using it for other productive methods," Chen said.