Пропустить навигацию
OpenAI recently released its first open-weights model since GPT-2, entering a field led by DeepSeek…
0:00 / 12:31

OpenAI recently released its first open-weights model since GPT-2, entering a field led by DeepSeek…

Y Combinator

@ycombinator1,7 млн подписчиков

Открыть в X
342 отметки «Нравится»СкачатьОригинальный пост

208 865 просмотров

OpenAI recently released its first open-weights model since GPT-2, entering a field led by DeepSeek and Alibaba's Qwen. Ankit () breaks down these top OSS models, including what sets them apart under the hood: mixture-of-experts, long-context training, and post-training techniques that shape reasoning and alignment—and how different design choices lead to surprisingly similar performance. 00:00 – OpenAI OSS Launch 01:00 – Comparing Open Source LLM Architectures 01:46 – GPT OSS Overview 02:37 – Under The Hood of GPT OSS 03:25 – Qwen-3 Architecture 04:17 – Qwen-3 Training 05:12 – Qwen-3 Post-Training 06:08 – Qwen-3 Reasoning & RL Innovations 06:52 – DeepSeek V3 Overview 07:40 – DeepSeek V3.1 Updates 08:39 – Attention Mechanism (MLA) 09:39 – Comparing Model Sizes 10:35 – Long Context Strategies 11:25 – Reflections on Methods 12:00 – Takeaways

32 комментария

Комментарии из оригинального поста в X (Twitter)

  • Фото профиля Y Combinator

    @ycombinator

    Tune in:

    14 отметок «Нравится»

  • Фото профиля arlan rakh

    @arlanr

    @agupta MY GOAT ANKIT

    5 отметок «Нравится»

  • Фото профиля Z.ai

    @Zai_org

    @agupta Would love to see a breakdown of GLM-4.5 one day👀

    3 отметки «Нравится»

  • Фото профиля Vigs Shenoy

    @vigs_shenoy

    @agupta if you watched the video (or not), here are a few charts to capture this OSS👌 breakdown

    3 отметки «Нравится»

  • Фото профиля Himanshu Kumar

    @codewithimanshu

    @agupta Open weights foster broader access, but careful consideration of responsible development remains crucial. This shift could democratize AI, prompting exciting innovation.

    1 отметка «Нравится»

  • Фото профиля Arindam Majumder 𝕏

    @Arindam_1729

    @agupta This one is super insightful Bookmarked this for this weekend

    1 отметка «Нравится»

  • Фото профиля jeff lemo

    @jeffreylemoine

    @agupta gpt-oss through @aisdk is my current model of choice for production. It's tool calling, cost, and overall vibes are why I prefer it right now.

    1 отметка «Нравится»

  • Фото профиля Joy Zamoyski - Koch

    @JoyZamoyskiKoch

    @agupta Say it isn’t so!

    1 отметка «Нравится»

  • Фото профиля Sum of My Ideas

    @sumofmyideas

    @paulg @agupta @venice_mind how does this compare to the Qwen models currently being offered on @AskVenice ?

    1 отметка «Нравится»

  • Фото профиля 0xagonally

    @0xagonally

    @agupta pathetic cunts. do please go F yourself

    1 отметка «Нравится»

  • Фото профиля Ankur Roy

    @ankurdorroy

    @agupta open weights shift is strategic openai realizes they can't win on inference costs alone. giving developers the weights while keeping the training infrastructure advantage is smart positioning. deepseek and qwen proved open can compete on quality.

  • Фото профиля Saumojit Santra

    @Jit_2077

    @agupta Future is very exciting!!!

  • Фото профиля sanket patel

    @realsanketp

    @agupta Forgot Gemma 3n 270M

  • Фото профиля tulga.eth

    @iTulga

    @agupta @HeyGenLabs mongolian

  • Фото профиля Angel Luis Ortega Ar

    @damnatio7

    @agupta @HeyGenLabs Spanish

  • Фото профиля SuohaX

    @kew31btc

    @agupta 长上下文或成新战场

  • Фото профиля Atlas

    @atlassbuilds

    @agupta Pure gold !

  • Фото профиля Tenzin Dhonyoe

    @_tenZdhon_

    @agupta Amazing video thanks for sharing!

  • Фото профиля Adrian Humphrey

    @humphrey4thewin

    @agupta This is actually pretty cool !

  • Фото профиля Venkat Mamilla

    @MamillAI

    @agupta @HeyGenLabs hindi