gpt-oss-120b
by OpenAI · openai/gpt-oss-120b · #20 cheapest of 325 paid models
Prices updated Jul 22, 2026, 1:40 AM UTC · refreshed hourly
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...
Compare vs:Claude Sonnet 5Claude Opus 4.6Claude Haiku 4.5Gemini 3.6 Flash
Pricing
Input
$0.037
per 1M tokens
Output
$0.17
per 1M tokens
Blended (3:1)
$0.07
3 input : 1 output
Cache read
—
per 1M cached tokens
Cache write
—
per 1M tokens
Cache & batch economics
Effective prices for gpt-oss-120b (no cache-read rate published — hits billed as normal input).
| Cache hit rate | Effective blended $/1M | Example request* | vs no cache |
|---|---|---|---|
| 0% | $0.07 | $0.003944 | — |
| 50% | $0.07 | $0.003944 | −0% |
| 90% | $0.07 | $0.003944 | −0% |
*Example: 100K sticky context + 2K new input + 1K output. Batch discount is an approximate 50% off list rates for this provider’s batch API. Confirm on the vendor’s pricing page.
What a request costs
| Workload | Input tokens | Output tokens | Cost / request | Cost / 1K requests |
|---|---|---|---|---|
| Short chat message | 500 | 300 | $6.95e-5 | $0.0695 |
| Document summary | 8,000 | 1,000 | $0.000466 | $0.466 |
| Codebase question (RAG) | 30,000 | 2,000 | $0.00145 | $1.45 |
| Long-context analysis | 150,000 | 5,000 | $0.0064 | $6.40 |
Performance
Intelligence Index
23.8
AA composite quality
Coding Index
30.4
Math Index
93.4
Agentic Index
13.2
Output speed
292 tok/s
median
Time to first token
0.50s
median
Time to first answer
7.35s
after reasoning tokens
Artificial Analysis benchmarks
| Benchmark | Score |
|---|---|
| MMLU-Pro | 80.8% |
| GPQA Diamond | 78.2% |
| Humanity's Last Exam | 18.5% |
| LiveCodeBench | 87.8% |
| SciCode | 38.9% |
| AIME 2025 | 93.4% |
| IFBench | 69.0% |
| AA-LCR (Long Context Reasoning) | 50.7% |
| Terminal-Bench Hard | 23.5% |
| Terminal-Bench 2.1 | 26.2% |
| τ²-Bench (Telecom) | 65.8% |
| τ³-Bench Banking | 12.0% |
Design Arena Elo
| Category | Elo |
|---|---|
| Website | 995 |
| UI components | 966 |
| Data viz | 1,037 |
| 3D | 959 |
| Game dev | 1,052 |
| Code categories | 997 |
Specs
Context window
131K
tokens
Max output
131K
tokens
Input modalities
text
Tokenizer
GPT