GPT OSS 20B API Pricing
As of October 10, 2026, GPT OSS 20B costs $0.0145 per 1M input tokens and $0.07 per 1M output tokens from Darkbloom, the lowest of 14 providers. It has a 131,072-token context window and returns up to 32,768 tokens per response.
- Input
- $0.0145 / 1M tokens
- Output
- $0.07 / 1M tokens
- Context window
- 131,072 tokens
- Max output
- 32,768 tokens
- Providers
- 14
- Added to LiteLLM
- August 2025
GPT OSS 20B pricing by provider
17 listings across 14 providers. Token prices are in US dollars per 1M tokens. The highlighted row is the price quoted above.
| Provider | Model name in LiteLLM | Input | Output | Cache read | Cache write | Batch in | Batch out | Context |
|---|---|---|---|---|---|---|---|---|
| Darkbloom | darkbloom/gpt-oss-20b | $0.0145 | $0.07 | – | – | – | – | 128K |
| Amazon Bedrock Mantle | bedrock_mantle/openai.gpt-oss-20b | $0.07 | $0.30 | – | – | – | – | 128K |
| Cloudflare Workers AI | cloudflare/@cf/openai/gpt-oss-20b | $0.20 | $0.30 | – | – | – | – | 128K |
| Databricks | databricks/databricks-gpt-oss-20b | $0.07 | $0.30 | $0.07 | $0.07 | – | – | 128K |
| DeepInfra | deepinfra/openai/gpt-oss-20b | $0.03 | $0.14 | – | – | – | – | 128K |
| Fireworks AI | fireworks_ai/accounts/fireworks/models/gpt-oss-20b | $0.07 | $0.30 | $0.035 | – | – | – | 128K |
| Fireworks AI | fireworks_ai/gpt-oss-20b | $0.07 | $0.30 | $0.035 | – | – | – | 128K |
| Google Vertex AI | vertex_ai/openai/gpt-oss-20b-maas | $0.07 | $0.25 | $0.007 | – | $0.035 | $0.125 | 128K |
| Groq | groq/openai/gpt-oss-20b | $0.075 | $0.30 | $0.0375 | – | – | – | 128K |
| Novita AI | novita/openai/gpt-oss-20b | $0.04 | $0.15 | – | – | – | – | 128K |
| OVHcloud | ovhcloud/gpt-oss-20b | $0.04 | $0.15 | – | – | – | – | 131K |
| Replicate | replicate/openai/gpt-oss-20b | $0.09 | $0.36 | – | – | – | – | – |
| Tensormesh | tensormesh/openai/gpt-oss-20b | $0.07 | $0.28 | $0 | – | – | – | 128K |
| Together AI | together_ai/openai/gpt-oss-20b | $0.05 | $0.20 | – | – | – | – | 128K |
| Weights & Biases | wandb/openai/gpt-oss-20b | $0.03 | $0.13 | – | – | – | – | 131K |
| Amazon Bedrock Mantleus-gov-east-1 | bedrock_mantle/us-gov-east-1/openai.gpt-oss-20b | $0.084 | $0.36 | – | – | – | – | 128K |
| Amazon Bedrock Mantleus-gov-west-1 | bedrock_mantle/us-gov-west-1/openai.gpt-oss-20b | $0.084 | $0.36 | – | – | – | – | 128K |
What GPT OSS 20B costs in practice
1,000 requests with 2,000 input and 500 output tokens each
| Darkbloom | $0.064 |
| DeepInfra | $0.13 |
| Weights & Biases | $0.125 |
| Novita AI | $0.155 |
100 long-document requests with 100,000 input and 2,000 output tokens each
| Darkbloom | $0.159 |
| DeepInfra | $0.328 |
| Weights & Biases | $0.326 |
| Novita AI | $0.43 |
GPT OSS 20B features
- Function calling
- Parallel tool calls
- Structured output
- Reasoning
- Prompt caching
- Web search
Use GPT OSS 20B with LiteLLM
Call GPT OSS 20B through the LiteLLM Python SDK, or put it behind the LiteLLM proxy and send OpenAI-format requests to /v1/chat/completions. LiteLLM tracks the cost of every request at these prices.
from litellm import completion
response = completion(
model="darkbloom/gpt-oss-20b",
messages=[{"role": "user", "content": "Hello!"}],
)model_list:
- model_name: gpt-oss-20b
litellm_params:
model: darkbloom/gpt-oss-20b
# credentials: see the provider's LiteLLM docs pagecurl http://0.0.0.0:4000/v1/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer sk-1234" \
-d '{
"model": "gpt-oss-20b",
"messages": [{"role": "user", "content": "Hello!"}]
}'GPT OSS 20B FAQ
How much does GPT OSS 20B cost?
GPT OSS 20B costs $0.0145 per 1M input tokens and $0.07 per 1M output tokens from Darkbloom, the lowest of 14 providers.
What is the context window of GPT OSS 20B?
GPT OSS 20B has a 131,072-token context window and returns up to 32,768 tokens per response.
Which providers offer GPT OSS 20B?
GPT OSS 20B is available from 14 providers through LiteLLM: Darkbloom, Amazon Bedrock Mantle, Cloudflare Workers AI, Databricks, DeepInfra, Fireworks AI, Google Vertex AI, Groq, Novita AI, OVHcloud, Replicate and Tensormesh and 2 more. The lowest price is on Darkbloom.
How do I call GPT OSS 20B with an OpenAI-compatible API?
Use the LiteLLM Python SDK with model="darkbloom/gpt-oss-20b", or add the model to the LiteLLM proxy's config.yaml and send requests to /v1/chat/completions from any OpenAI client. LiteLLM tracks the cost of every request at the prices on this page.
What features does GPT OSS 20B support?
GPT OSS 20B supports function calling, parallel tool calls, structured output, reasoning, prompt caching and web search.
When did LiteLLM add GPT OSS 20B?
GPT OSS 20B was added to LiteLLM's model price file in August 2025.
Sources
Prices come from LiteLLM's open-source model_prices_and_context_window.json, which LiteLLM uses to calculate the cost of every request. Provider pricing pages:
- www.darkbloom.dev/
- docs.aws.amazon.com/bedrock/latest/userguide/model-card-openai-gpt-oss-20b.html
- www.databricks.com/product/pricing/foundation-model-serving
- deepinfra.com/pricing
LiteLLM docs: Darkbloom, Amazon Bedrock Mantle, Cloudflare Workers AI, Databricks, DeepInfra, Fireworks AI.
Provider pages: Darkbloom, Amazon Bedrock Mantle, Cloudflare Workers AI, Databricks, DeepInfra, Fireworks AI, Google Vertex AI, Groq, Novita AI, OVHcloud, Replicate, Tensormesh.