gpt-oss-120b
OpenAI🇺🇸 United States
OpenAI's first open-weights model since GPT-2. 117B-param MoE (5.1B active) under Apache 2.0, near o4-mini performance, runs on a single 80GB GPU. Configurable reasoning effort.
Context window131K tokens
Input / 1M tokens$0.037
Output / 1M tokens$0.17
Version History
gpt-oss-120b-launchmajor
gpt-oss-120b launches. OpenAI's first open-weights model since GPT-2. 117B-param MoE (5.1B active) under Apache 2.0, near o4-mini performance, runs on a single 80GB GPU. Configurable reasoning effort.
Benchmark Scores
Full leaderboard →93.4%
AIME 2025
1352.0 elo
Arena Elo
78.2%
GPQA
14.2%
Hallucination (Summarization)
90.8%
Hallucination Rate
87.8%
LiveCodeBench
90.0%
MMLU
80.8%
MMLU-Pro
201.0 tokens_per_sec
Speed (tok/s)