gpt-oss

“I would be very surprised if the original drop is anything more than totally obsolete/irrelevant at this point”

“I still think that was a really great model that got overlooked. It was really great in terms of latency/throughput while still being fairly intelligent.”

“Gemma 4 (April), Llama (April), and gpt-oss (August 2025) are up to date, but old (for models).”

“The OSS models are good, but weren't very competitive at release with alternatives like Qwen3 Coder A3B.”

“you can get away with models like gpt-oss that you can get from inference providers providing 10x+ the token/second speed”

“I've used it to suggest itineraries for trips and to review the completeness of small requests for proposals”

“What's wrong with a country hosting a Kimi, Qwen or GPT-Oss on their hardware for their government work purpose?”

“The gpt-oss models are good but there's no evidence that OpenAI have ongoing development on open models.”

“there are better tools with ggml-org/gpt-oss-20b-GGUF where you can see a less terse refusal for the prompt”

“I needed many more prompts for GPT OSS 20B to complete my contrived task, so progress felt much slower.”

“gpt-oss-120b: (unknown prefill), ~175 tok/s generation. I don't remember the prefill speed but it certainly was below 10k”

“gpt-oss is hilariously stupid, but might work for typescript, react, python simple tasks”

“OSS 120B on Groq/Cerebras is one of the fastest models available with enough smarts for basic tasks”

“Especially with something like gpt-oss-20b, where I've found executing tools happens at a higher success than claude sonnet via openrouter.”

“it started functioning extremely well, and it's the main model I use for my homemade agents”

“Both models with the reasoning_effort set to high are actually quite good compared to other downloadable models”

“offloading tasks to smaller models (gpt-oss specifically for me) gets data to the bigger model quicker”

“I’ve found it to be incredibly weak in the real world every single time I’ve reached for it.”

“pretty great tbh - one of the better all-around local models for knowledge and grounding”

“which according to my own benchmarks seems to still be the strongest local model I can run myself”

“open models like Google's gemma and OpenAI's gpt-oss have behaviors more similar to their paid counterparts”

“My local GPT-OSS 120B is fantastic at generating the scene, descriptors for the environment”

“It's a really good model for its size, but context length is a serious issue to avoid hallucination”

“providers on OpenRouter make it essentially free and instant, and it's roughly as capable at o4-mini was in my experience”

“runs at perfectly servicable pace locally on a laptop 5090 with 64gb system ram with zero effort required”

“If you have a bunch of money and patience, you can also run something like GPT OSS 120B or GLM 4.5 Air locally.”

“I find it quite decent for a lot of things, especially after I set reasoning effort to high”

“In my experience it's output is comparable to what you'd get out of older models and the openAI benchmarks seem accurate”

“Once I did that it started functioning extremely well, and it's the main model I use for my homemade agents.”

“gpt-oss struggled and RAG benchmarks suddenly went from 90% to 60% without additional prompt engineering”

“Cerebras gpt-oss-120b for post-processing (cleaning up transcription errors and fixing technical mondegreens”

“It was so heavily censored and lobotomized that it has become a meme in the local LLM community.”

“will run gpt-oss-120b at a very respectable clip and I've found it to be quite serviceable for a lot of tasks”

“due to its ultra low latency often completing full generations before Gemini or ChatGPT APIs return even the first token”

“It's a decent general model too - I have it plugged up in llm and raycast since August at great speeds.”

“I am using gpt-oss for a project right now, I think its quality is comparable to the ones I mentioned.”

“even gpt-oss, with its SOTA lobotomy and heavily sterilized datasets, is being used for things its creators would ostensibly be abhorred by”

“Some models cost almost nothing like GPT OSS 120B which can be used to create slop at the speed of light.”

“for even moderately difficult programming tasks, 20B fails repeatedly while 120B gets it right on the first try”

“20B works great for small-ish text summarization and title generation”