Loading video…
All breakouts
DeepSeekseen in 18 videos →
13.1Talking HeadContrarian hook
Hook
Local models will win
@lord.luke5
1.9K followers · usually 473 views a post
Views
6.0K
Likes
211
Comments
58
Shares
10
Selling · SaaS › AI Tool
Why it worked
The video leverages a compelling data point (cost-effectiveness of AI models) and presents it in a visually engaging way with on-screen text, making a strong argument for the superiority of local models like DeepSeek.
Summary
The video argues that local AI models will win due to their efficiency and cost-effectiveness, highlighting DeepSeek's performance compared to other models. The creator encourages viewers to join their community to learn more about leveraging these models.
Structure
- 1Introduction: Local models will win
- 2Comparison: DeepSeek outperforms Fable on cost and performance
- 3Argument: Efficiency is key to winning
- 4Call to action: Join the community to learn more
Product placement
The video displays a benchmark chart comparing AI models, with DeepSeek V4-Flash highlighted for its performance and low cost. The creator discusses DeepSeek's efficiency and cost-effectiveness.
Call to action
Please join the community if you want to learn more on how to take advantage of this.
On-screen text
Local models will win
Terminal-Bench 2.1
Latest scores, Jul 2026 - price per 1M tokens (in/out)
85.8
82.7
80.5
74.5
GPT-S6 Sol
$5 / $30
DeepSeek V4-Flash
$0.14 / $0.28
Fable 5
$10 / $50
Sonnet S
$3 / $15
IT
ABSURD
COST
UNDER
DOLLAR
OUTPERFORMING
FABLE
ON
BENCH.
BE
PRO
CLOUD,
CHAT
PERSON
JUST
BE
RIDICULOUS.
BECAUSE
MAKING
EFFICIENT.
CRAZY
THAT
MUCH
DISAGREEMENT,
WHEN
THIS.
PLEASE
COMMUNITY
MORE
TO
THIS.
Original caption
How is something that’s under $.50 outperforming fable this is just so ridiculous#greenscreen #ai #Tech #deepseek
More from @lord.luke5
Their other posts in the index, biggest breakout first.
