gpt-oss-20b (free)
FREEby OpenAI · openai/gpt-oss-20b:free
Prices updated Jul 22, 2026, 1:36 AM UTC · refreshed hourly
gpt-oss-20b is an open-weight 21B parameter model released by OpenAI under the Apache 2.0 license. It uses a Mixture-of-Experts (MoE) architecture with 3.6B active parameters per forward pass, optimized for...
Compare vs:Claude Sonnet 5Claude Opus 4.6Claude Haiku 4.5Gemini 3.6 Flash
Pricing
Input
Free
per 1M tokens
Output
Free
per 1M tokens
Blended (3:1)
Free
3 input : 1 output
Cache read
—
per 1M cached tokens
Cache write
—
per 1M tokens
Cache & batch economics
Effective prices for gpt-oss-20b (free) (no cache-read rate published — hits billed as normal input).
| Cache hit rate | Effective blended $/1M | Example request* | vs no cache |
|---|---|---|---|
| 0% | Free | $0 | — |
| 50% | Free | $0 | −0% |
| 90% | Free | $0 | −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 | $0 | $0 |
| Document summary | 8,000 | 1,000 | $0 | $0 |
| Codebase question (RAG) | 30,000 | 2,000 | $0 | $0 |
| Long-context analysis | 150,000 | 5,000 | $0 | $0 |
Performance
Intelligence Index
14.9
AA composite quality
Coding Index
20.7
Math Index
89.3
Agentic Index
3.1
Output speed
253 tok/s
median
Time to first token
0.39s
median
Time to first answer
8.29s
after reasoning tokens
Artificial Analysis benchmarks
| Benchmark | Score |
|---|---|
| MMLU-Pro | 74.8% |
| GPQA Diamond | 68.8% |
| Humanity's Last Exam | 9.8% |
| LiveCodeBench | 77.7% |
| SciCode | 34.4% |
| AIME 2025 | 89.3% |
| IFBench | 65.1% |
| AA-LCR (Long Context Reasoning) | 30.7% |
| Terminal-Bench Hard | 10.6% |
| Terminal-Bench 2.1 | 13.9% |
| τ²-Bench (Telecom) | 60.2% |
| τ³-Bench Banking | 7.0% |
Design Arena Elo
| Category | Elo |
|---|---|
| Website | 880 |
| Data viz | 971 |
Specs
Context window
131K
tokens
Max output
33K
tokens
Input modalities
text
Tokenizer
GPT