Tremendous

An angel investor's take on life and business

Tencent is worth $600 billion, making it the largest company in China. But very few Americans have tried its AI models. This morning, I tested their most powerful model yet…

Tencent released Hunyuan-T1 in April. T1 is their most powerful reasoning model. But can it beat the best from America?

Let’s find out!

Round #1: Finding Missile Defense Startups

I read Nuclear War: A Scenario last year and it scared the heck out of me. Ever since, I’ve been looking for missile defense startups.

Let’s see what T1 can find…

T1 gave me a wonky response. It found only one startup, Black Swift Technologies. Confusingly, it marked it as number two, but there was no number one. What’s more, the link it provided as a citation was broken.

When I looked up Black Swift, I found it had nothing to do with missile defense. It produces drones for scientific and industrial applications.

This response was utterly useless — I’m giving it an F.

Round #2: Beating the Heat

Let’s give T1 something a little simpler but still useful…

Lately, it’s been hotter than death here in North Jersey. What are some great ways to beat the heat?

T1 did much better here, giving me some interesting ideas like avoiding spicy food. Spicy food can make you sweat, which isn’t very helpful in humid climates like mine.

However, it didn’t cite any sources in its response, even though I told it to base its answer on research. Citations are table stakes these days — any decent AI model can do it. There’s no excuse for T1 missing that.

Still, T1 provided me some useful info. I’ll give it a B- for this round.

Round #3: Teach Me About Solar Energy

I want to see a future in which we use way more power than ever. But all our power comes from the sky, and neither the Russians nor the Arabs can take it away.

So for our final round, I asked T1 about the growth of solar energy in America. How much of our energy is coming from the sun today, and how quickly is that number growing?

T1’s response looked impressive, with exact figures and a source. But the source was dated — a 2023 report.

When I looked at the report, the numbers didn’t line up with T1’s response. T1 was off by around 20%.

T1 gave a roughly correct response, but its sourcing was weak. I’ll give it a C here.

Wrap-Up

T1 barely survived my testing, notching a C- overall.

In all, I found Hunyuan-T1 to be an embarrassingly poor AI model. It does well on benchmarks, but benchmarks can be gamed by “teaching to the test.” Perhaps that’s what Tencent did here.

Tencent is way behind the leaders in China like Moonshot’s Kimi and DeepSeek. And it’s even further behind Grok.

If Tencent wants to stay on top, it’s going to have to do a lot better than this.

More on tech:

Grok Companions — Elon’s AI Girlfriend?

Grok 4: The Best AI Model Ever?

Where I’m Most Excited About Investing Now

Save Money on Stuff I Use:

Fundrise

This platform lets me diversify my real estate investments so I’m not too exposed to any one market. I’ve invested since 2018 with great returns.

More on Fundrise in this post.

If you decide to invest in Fundrise, you can use this link to get $100 in free bonus shares!

Misfits Market

I’ve used Misfits for years, and it never disappoints! Every fruit and vegetable is organic, super fresh, and packed with flavor!

I wrote a detailed review of Misfits here.

Use this link to sign up and you’ll save $15 on your first order. 

One response to “Can China’s Top Tech Company Beat America in AI? — Testing Hunyuan-T1”

  1. […] Can China’s Top Tech Company Beat America in AI? — Testing Hunyuan-T1 […]

    Like

Leave a comment