CHINESE AI

@malikontech

2.0K followers · usually 1.2K views a post

Views

7.5K

Likes

174

Comments

21

Shares

13

Selling · SaaS › AI Tool

DeepSeekseen in 18 videos →https://deepseek.com/quick_start/api ↗

Why it worked

The video leverages a controversial and attention-grabbing claim (Chinese AI models surpassing US models) and presents compelling data (cost, size, performance benchmarks) to support its argument. The comparison to well-known models like Opus and the discussion of open-source vs. closed-source AI likely resonated with the target audience interested in AI advancements.

Summary

The video discusses the release of DeepSeek-V4-Flash, an AI model that claims to be significantly cheaper and smaller than competitors like Opus, while offering comparable or superior performance. The speaker highlights that open-source AI models are rapidly advancing and may soon surpass closed-source models.

Structure

1Introduction of DeepSeek-V4-Flash API
2Comparison to Opus and cost/size advantages
3Discussion of open-source vs. closed-source AI
4Benchmarking and performance claims
5Prediction of open-source AI dominance

Original caption

Is this the turning point for AI globally that will force more companies to opensource? #ai #deepseek #china #claude

More from @malikontech

Their other posts in the index, biggest breakout first.

Product placement

The video features a man discussing the DeepSeek-V4-Flash API, with on-screen text and a table comparing benchmark scores of various AI models, including DeepSeek's offerings against GLM-5.2 and Opus 4.8. The DeepSeek website URL is also displayed.

Call to action

Check out the configuration details in our official API docs: deepseek.com/quick_start/api

On-screen text

DeepSeek-V4-Flash Official API is now LIVE in public beta! BETTER THAN OPUS? We've massively upgraded its Agent capabilities—benchmark scores are now far surpassing the V4-Pro-Preview. Check out the massive performance leap below! The official V4-Flash now natively supports the Responses API format and is fully adapted for Codex! Check out the configuration details in our official API docs: deepseek.com/quick_start/api

Transcript

CHINESE AI models are getting better than US. models, and it's clear and obvious at this point. DeepSeek V4 just came out and it is 100 times cheaper than Opus and it's 10 times smaller. I don't know how they're doing this. It's actually insane at this point that they're pulling this off. I have not used it intelligence wise. Obviously, these are benchmarks, you know, benchmarks. We don't pay attention to them. But the problem is, is that this is going to continue with closed frontier laps. When you have closed, when you have open AI, when you have Grot, etcetera, they are open source. This is just proof that that does not matter. Open source companies like DeepSeek, Kimmy, etcetera are benchmarking against the best closed source models, and they are so close that they might just win the AI race. And it might be happening in front of our eyes.