gpt-oss-20b pricing
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 lower-latency inference and deployability on consumer or single-GPU hardware. The model is trained in OpenAI’s Harmony response format and supports reasoning level configuration, fine-tuning, and agentic capabilities including function calling, tool use, and structured outputs. This page tracks 9 listings in total. Highlighted lows are $0.020 per million input and $0.100 per million output (see table for which seller matches each).
Pricing across providers
All figures are list prices per million tokens unless a column says otherwise. 9 offers are listed for gpt-oss-20b. Best input in this view: Openrouter.
| Provider | Input / 1M | Output / 1M | Cached input | Batch |
|---|---|---|---|---|
O Openrouter | $0.020 | $0.100 | — | — |
W Wandb | $5000.00 | $20000.00 | — | — |
O Ovhcloud | $0.040 | $0.150 | — | — |
N Novita | $0.040 | $0.150 | — | — |
D DeepInfra | $0.040 | $0.150 | — | — |
R Replicate | $0.090 | $0.360 | — | — |
FA Fireworks AI | $0.050 | $0.200 | — | — |
TA Together AI | $0.050 | $0.200 | — | — |
G Groq | $0.075 | $0.300 | $0.037 | — |
Input vs output · per provider
Cost calculator
Use this block to stress test gpt-oss-20b cost without a spreadsheet. All estimates come from public list rates in this page.
Provider
0.002000¢ / req
0.005000¢ / req
Model specifications
These fields describe gpt-oss-20b as we store it (source: OpenAI). They sit next to price so buyers can check limits and tools in one place.
- Context window
- 131,072 tokens
- Max output
- 131,072 tokens
- Vision (images)
- No
- Tool / function calling
- Yes
- Streaming
- No
- Released
- Aug 2025
- Primary provider
- OpenAI
- Model family
- N/A
Compare gpt-oss-20b
Open a pair page to see gpt-oss-20b next to another model with a shared provider matrix. 6 shortcuts below.
- gpt-oss-20b vs GPT-4o
Compare pricing side by side
- gpt-oss-20b vs GPT-4o mini
Compare pricing side by side
- gpt-oss-20b vs o3
Compare pricing side by side
- gpt-oss-20b vs Claude Sonnet 4.6
Compare pricing side by side
- gpt-oss-20b vs Gemini 2.0 Flash
Compare pricing side by side
- gpt-oss-20b vs Llama 3.1 70B
Compare pricing side by side
Frequently asked questions
Answers pull from the same numbers you see on this page. The short model note from our index: 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 lower-latency...
Also from OpenAI
Other models by OpenAI with live pricing in our catalog.