↑ ↓ Navigate ↵ Select ESC Close

Why is OpenAI so much more efficient?

2026-06-30 Science & Technology
5.1k
287
38
Theo - t3․gg
Theo - t3․gg
552.0k subscribers

Unlock all features

FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.

Get My 10 Free Today

Related videos

This is bad...
27:17

This is bad...

Theo - t3․gg

104.3k views

Microsoft and OpenAI break up (Amazon is pumped)
36:46

Microsoft and OpenAI break up (Amazon is pumped)

Theo - t3․gg

112.2k views

Open source is dead now?
26:21

Open source is dead now?

Theo - t3․gg

82.2k views

OpenAI is lying
38:58

OpenAI is lying

Theo - t3․gg

81.9k views

Open source is dying
40:27

Open source is dying

Theo - t3․gg

97.8k views

Never mind (OpenAI won again)
43:33

Never mind (OpenAI won again)

Theo - t3․gg

97.8k views

Anthropic just burned so much trust...
24:25

Anthropic just burned so much trust...

Theo - t3․gg

159.2k views

OpenAI: Trapped in 2nd place
33:55

OpenAI: Trapped in 2nd place

Theo - t3․gg

59.0k views

I’m so tired.
20:47

I’m so tired.

Theo - t3․gg

65.4k views

OpenAI is in "CODE RED" (Did Gemini win that hard??)
32:39

OpenAI is in "CODE RED" (Did Gemini win that hard??)

Theo - t3․gg

72.0k views

Description

OpenAI's models traditionally score higher on coding benchmarks using a FRACTION of the tokens of other models, especially Gemini's 250k per task average, and that's the benefit of the "Grug speak" theory of model efficiency that was leaked in certain GPT 5.5 reasoning traces. Ya'll said you missed technical deep dives so here's one for you nerds. Thank you Browserbase for sponsoring! Check them out at: https://soydev.link/browserbase Want to sponsor a video? Learn more here: https://soydev.link/sponsor-me Check out my Twitch, Twitter, Discord more at https://t3.gg S/O @Ph4seon3 for the awesome edit 🙏 #ai #coding #programming

Top Comments (10)

@justofon 2026-06-30

Here before the inevitable obsidian comment

318 7 replies
@josikinz 2026-06-30

Big linguistics doesn’t want you to know that Chinese caveman speech is the most token efficient form of communication 😤

293 30 replies
@juliusbrussee 2026-06-30

So you're telling me, OpenAI uses Caveman trained reasoning to increase the quality and efficiency of their models...

235 6 replies
@PurpleDaemon_ 2026-06-30

0:58 the most insane here is right to left graph

221 14 replies
@BeInnovativeToday 2026-06-30

Thank you we need to push the TOKEN EFFICIENCY Movement

72 4 replies
@alecwip 2026-06-30

hopefully gpt5.6 before the end of the year 😢

52 7 replies
@olli3008 2026-06-30

Claude's reasoning traces are still available on their web chat.

40 10 replies
@debbanerji7301 2026-06-30

Smart computer man teach fast word use

25
@petereklund7145 2026-06-30

lol, those reasoning traces are basically how I write when I write temporary working notes for myself that I expect no other people to read. Especially if I write with pen and paper.

18
@noblerawa 2026-06-30

The only YT channel i dont skip sponsors on since its actually useful

3

Unlock the Data Inside
Turn Videos into Knowledge

  • Get FREE 10/day: transcripts, summaries, chats
  • Chat with videos, export text & PDF
  • $1 free API credit for RAG, chatbots & research
Try it free

Free forever plan • All features unlocked

App screenshot