# Requests to open source additional OpenAI models

**URL:** <https://community.openai.com/t/requests-to-open-source-additional-openai-models/1381698>\
**Category:** Open Models\
**Tags:** open-source, gpt-4o, gpt-oss-20b\
**Created:** [May 24, 2026, 7:27pm UTC](https://community.openai.com/t/requests-to-open-source-additional-openai-models/1381698 "2026-05-24T19:27:38Z")\
**Posts on this page:** 4\
**Page:** 1

<div class="post-metadata">

**Author:** ![Erik\_AI\_22](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/erik_ai_22/32/723341_2.png) [@Erik\_AI\_22](https://community.openai.com/u/Erik_AI_22)\
**Post date:** [May 24, 2026, 7:27pm UTC](https://community.openai.com/t/requests-to-open-source-additional-openai-models/1381698/1 "2026-05-24T19:27:38Z")

</div>

OpenAI should release GPT-OSS-48B, with 48B total parameters and 5B active.

It should be distilled from GPT-5.5 with reasoning effort high (and medium in some cases in data).

Should be trained on a filtered Common Crawl CC-MAIN-2026-17 and a lot of more data. So it’s knowledge cutoff would be April 23, 2026 (Note: Teacher has a knowledge cutoff of December 1, 2025.)

OpenAI should pretrain GPT-OSS-2-48B on 24T tokens.

OpenAI should support 96 languages for the model (32 core languages: high-quality chat, reasoning, coding help, summaries, 32 secondary languages: usable translation, QA, simple chat, retrieval-style tasks, 32 programming languages)

**32 core languages**

1. English

2. Spanish

3. French

4. German

5. Italian

6. Portuguese

7. Dutch

8. Polish

9. Russian

10. Ukrainian

11. Turkish

12. Arabic

13. Hebrew

14. Hindi

15. Bengali

16. Urdu

17. Persian

18. Indonesian

19. Malay

20. Vietnamese

21. Thai

22. Chinese

23. Japanese

24. Korean

25. Catalan

26. Basque

27. Swedish

28. Norwegian

29. Danish

30. Finnish

31. Greek

32. Czech

## **32 secondary supported languages**

1. Slovak

2. Slovenian

3. Croatian

4. Serbian

5. Bulgarian

6. Romanian

7. Hungarian

8. Lithuanian

9. Latvian

10. Estonian

11. Irish

12. Welsh

13. Icelandic

14. Albanian

15. Macedonian

16. Georgian

17. Armenian

18. Azerbaijani

19. Kazakh

20. Uzbek

21. Swahili

22. Yoruba

23. Hausa

24. Afrikaans

25. Amharic

26. Tamil

27. Telugu

28. Marathi

29. Gujarati

30. Kannada

31. Malayalam

32. Tagalog

## **32 supported programming languages**

1. Python

2. JavaScript

3. TypeScript

4. HTML

5. CSS

6. SQL

7. Bash / Shell

8. C

9. C++

10. C#

11. Java

12. Kotlin

13. Swift

14. Go

15. Rust

16. PHP

17. Ruby

18. R

19. Lua

20. Dart

21. Scala

22. Objective-C

23. Perl

24. PowerShell

25. MATLAB

26. Julia

27. Haskell

28. Elixir

29. Erlang

30. Zig

31. Assembly

32. Fortran

---

<div class="post-metadata">

**Author:** ![Erik\_AI\_22](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/erik_ai_22/32/723341_2.png) [@Erik\_AI\_22](https://community.openai.com/u/Erik_AI_22)\
**Post date:** [May 25, 2026, 1:29pm UTC](https://community.openai.com/t/requests-to-open-source-additional-openai-models/1381698/2 "2026-05-25T13:29:48Z")

</div>

Dear OpenAI Developers, respond with a emoji cuz’ Yes or No. People are angry because openai deleted back in February 13 GPt 4o, and it was the first human and emotional model. so wee need to open source it and upload it to hugging face. Please. See this now, Sam Altman! It is ~200B and MoE params, same as DeepSeek V4 flash, minimax m2.7. Also open source gpt 4o mini so on Consumer hardware can run it Bc it is ~40B With ~ 8B activated! #PleaseOpenSource4o

---

<div class="post-metadata">

**Author:** ![Erik\_AI\_22](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/erik_ai_22/32/723341_2.png) [@Erik\_AI\_22](https://community.openai.com/u/Erik_AI_22)\
**Post date:** [May 25, 2026, 3:54pm UTC](https://community.openai.com/t/requests-to-open-source-additional-openai-models/1381698/4 "2026-05-25T15:54:58Z")

</div>

Want GPT-OSS-2 since no open weights model since 08/2025 (August 2025). Model details. Model name: GPT-OSS-2-14B. Parameters: Total: 14B. Active: 5B. Knowledge cutoff: 06/2025 (June 2025). Training: Trained on diverse and multilingual data up to 06/2025 (June 2025), 15T-25T tokens. Upload it to hugging face, on OpenAI org, also avalible on chatgpt on free tier after the 10 free gpt 5.5 instant messages. “openai/gpt-oss-2-14b” — Hugging Face Repo ID. If any OpenAI team guy is seeing, this, please create it

---

<div class="post-metadata">

**Author:** ![Sunn\_e](https://avatars.discourse-cdn.com/v4/letter/s/b5ac83/32.png) [@Sunn\_e](https://community.openai.com/u/Sunn_e)\
**Post date:** [May 27, 2026, 5:47pm UTC](https://community.openai.com/t/requests-to-open-source-additional-openai-models/1381698/5 "2026-05-27T17:47:03Z")

</div>