Hub/Models/openai/gpt-oss-120b
openaitext → text

OpenAI: gpt-oss-120b

Canonical ID: openai/gpt-oss-120b

Compare ModelOpen on OpenRouter →

gpt-oss-120b is an open-weight, 117B-parameter Mixture-of-Experts (MoE) language model from OpenAI designed for high-reasoning, agentic, and general-purpose production use cases. It activates 5.1B parameters per forward pass and is optimized...

Artificial Analysis Benchmarks

Verified
Intelligence Index3.8 / 100
Engineering Index11.2 / 100
Strategy & Ops5.5 / 100
Finance & Accounting7.8 / 100
Legal Index8.1 / 100
Healthcare & Medical5.8 / 100

Token Pricing & Economics

Prompt Input (1M Tokens)
$0.09
Input
Completion Output (1M Tokens)
$0.36
Output
Prompt KV-Cache Read (1M Tokens)
$0.00
Cached
Context Window Capacity
128k tokens
Context

Supported Parameters & Features

frequency_penaltyinclude_reasoninglogit_biaslogprobsmax_tokensmin_ppresence_penaltyreasoningreasoning_effortrepetition_penaltyresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_atop_ktop_logprobstop_p