Model specification · OpenAI

gpt-oss-120b

openai/gpt-oss-120b

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...

SpecificationsReported values
Context window
131K
Maximum output
131K
Architecture
text->text
Tokenizer
GPT
Knowledge cutoff
2024-06-30
Moderated
No
Input modalities
text
Output modalities
text
Full pricingUSD where reported
PriceStandard
Input$0.037 / M
Output$0.170 / M
Benchmark indexesReported indexes
Intelligence
23.8
Coding
30.4
Agentic
13.2
Design Arena Elo
996
Example workload costsStandard pricing only
WorkloadInput / outputCache shareEstimated cost
Quick check10K / 1K × 10%$0.00054
Long analysis100K / 4K × 150%$0.00438
Workday20K / 2K × 1,00030%$1.08
More from OpenAIAll OpenAI models
ModelContextInput / MOutput / MIntel.
GPT-5.6 Sol1.1M$5.00$30.0058.9
GPT-5.6 Terra1.1M$1.00$6.0055
GPT-5.51.1M$5.00$30.0054.8
GPT-5.41.1M$2.50$15.0051.4
GPT-5.6 Luna1.1M$0.100$0.60051.2
GPT-5.4 Mini400K$0.750$4.5040
GPT-5.4 Nano400K$0.200$1.2538.2
GPT-5.1400K$1.25$10.0036.9
GPT-5400K$1.25$10.0034.7
GPT-5 Mini400K$0.250$2.0025.3
o3 Mini High200K$1.10$4.4015.6
gpt-oss-20b131K$0.030$0.13014.9