openai/gpt-oss-120b · Discussions

#100 opened 4 months ago by

wasiuddina

How to reproduce the results on hle(Humanity's Last Exam)?

#95 opened 4 months ago by

wenhanli

ImportError: /lib64/libc.so.6: version `GLIBC_2.32' not found

#86 opened 4 months ago by

yueqiren

Why is the performance worse on the 20B than on the 120B

➕ 3

#84 opened 4 months ago by

megabob

How to turn off thinking mode

#82 opened 4 months ago by

Gierry

gpt-oss120b does not actually support 131072 output tokens due to openai policies embedded limiting ouput

8

#81 opened 4 months ago by

Theodophilus

Run GPT-OSS-120B with just Single A100 (80GB)

#80 opened 4 months ago by

ghostplant

Please move metal/ and original/ to their own repos

➕ 2

#78 opened 4 months ago by

ehartford

gp-oss-120b — Exceptional Reasoning, Not Yet AGI Scale

❤️ 3

#77 opened 4 months ago by

BertrandCabotPro

Qwen3 beat gpt-oss with just 0.6B with good quality enough to be usable

#75 opened 4 months ago by

yousef1727

Hallucinates System-promp

#74 opened 4 months ago by

Gradois

Can gpt-oss support local vllm deployment on a100 GPU?

10

#73 opened 4 months ago by

Cola-any

gpt_oss error when running on kaggle

#72 opened 4 months ago by

thehai

is this guid prompt is still valid ?

#71 opened 4 months ago by

gopi87

[Discussion] gpt-oss-120b hangs indefinitely ("thinking...") when using YaRN rope scaling to extend context length

#70 opened 4 months ago by

RekklesAI

Errors in chat template compared to spec

14

#69 opened 4 months ago by

zhuexe

Possible PEP 660 violation in `_build/gpt_oss_build_backend/backend.py`

#66 opened 4 months ago by

kwojciechowski

Estimated Thinking/Reasoning Token Usages For Each Modes

#65 opened 4 months ago by

asif00

running mxfp4 on H100 using tranformers with triton_kernel: make_default_matmul_mxfp4_w_layout not found

#64 opened 4 months ago by

uillliu

🚀 Best Practices for Evaluating GPT-OSS Models: Speed & Benchmark Testing Guide

🚀 ❤️ 11

#62 opened 4 months ago by

Yunxz

How do I serve a model in the original folder as bf16 in VLLM?

4

#60 opened 4 months ago by

bakch92

Model Performance

😔 🤗 6

#59 opened 4 months ago by

Joe1998

Disgusting, maximally censored model!

👍 35

17

#56 opened 4 months ago by

Lord-Kvento

🔥🔥🔥中文测评视频

#55 opened 4 months ago by

leo009

Llama, Mistral, Gemma… and now OpenAI enters the hunger games. 🐎⚔️

🤯 🔥 2

#54 opened 4 months ago by

Stephen555

Request: FP8 / BF16 version of model?

👍 3

#53 opened 4 months ago by

Epliz

кал

➕ 👍 3

#51 opened 4 months ago by

Serg4451D

how to disable the reasoning mode?

👍 12

7

#50 opened 4 months ago by

szzzzz

Enterprise AI factory OS

#47 opened 4 months ago by

DavidSteinbauer

How to use different reasoning effort in the example?

👍 🔥 2

#45 opened 4 months ago by

TianheWu

Structured Outputs

#44 opened 4 months ago by

DykeF

Problems with Metal

#42 opened 4 months ago by

Thalesian

[v1 engine][flash_attn backend] TypeError: flash_attn_varlen_func() got an unexpected keyword argument 's_aux' when running gpt-oss-120b on H200

👍 8

14

#41 opened 4 months ago by

RekklesAI

openai_harmony.HarmonyError: error downloading or loading vocab file: failed to download or load vocab file

11

#39 opened 4 months ago by

rsullenbLL

Cooooool !大模型本地部署，能力评测，幻觉评测微信交流群，欢迎感兴趣的朋友加入

#38 opened 4 months ago by

jakyer

I feel unsafe

🤗 5

9

#37 opened 4 months ago by

lmganon123

Deploy gpt-oss models in your own AWS account using vLLM and Tensorfuse

🔥 2

#36 opened 4 months ago by

agam30

Local Installation Video and Testing - Step by Step

#34 opened 4 months ago by

fahdmirzac

vLLM FlashAttention3 with A6000

👍 16

19

#33 opened 4 months ago by

YieumYoon

Request: 4-bit GPTQ or AWQ quantized version of openai/gpt-oss-20b

🚀 ❤️ 14

12

#32 opened 4 months ago by

powtac

Benchmaxed with no world knowledge or intuition. (ex. Wheelies on a motorcycle do not use front brakes to adjust the height b/c the front wheel is already off the ground duh)

👍 8

#31 opened 4 months ago by

TroyDoesAI

Streaming Issue

#29 opened 4 months ago by

seanliu96

Model claims to be GPT4-Turbo - any explanation behind this?

#28 opened 4 months ago by

marksverdhei

Web browsing (using built-in browsing tools)

#27 opened 4 months ago by

eryk-mazus

Knowledge limitations

👍 2

#25 opened 4 months ago by

hexess

VLLM - Flash-attn 3

12

#23 opened 4 months ago by

chriswritescode

From 🤗 to 🤯 — The Evolution

❤️ 🔥 9

#22 opened 4 months ago by

gokularaman

Recommended Sampling Parameters

👍 3