MODEL / OPENAI
gpt-oss-120b
OpenAI targets one 80 GB accelerator. Correct serving requires Harmony format support.
Fonte primáriaParâmetros totais117B5.1B active / token
Memória estimada~80 GBMXFP4
Contexto128KText
LicenseApache-2.0Primarily English
01 / IDEAL PARA
- reasoning
- agents
- datacenter serving
02 / AJUSTE ESTIMADO
03 / RODAR COM
vLLMAdvanced
vllm serve openai/gpt-oss-120b --dtype autoTensorRT-LLMAdvanced
trtllm-serve openai/gpt-oss-120bllama.cppMedium
llama-server -m /path/to/verified-model.gguf --port 8080Atenção
OpenAI targets one 80 GB accelerator. Correct serving requires Harmony format support.