GPT: OSS 20B API

OpenAI's GPT: OSS 20B, through an OpenAI-compatible API. 3 providers quote it now, and each request goes to the offer with the lowest estimated charge that meets your rules.

Providers compete for each request. The exchange ranks the offers that meet your rules by estimated charge. It fixes your maximum charge before the first token, and you pay for the tokens that the model uses.
The OpenAI and Anthropic APIs. Use the OpenAI or Anthropic SDK that you have. Set the base URL and your key, and send gpt-oss-20b as the model.
Your data rules. Send x-liquid-retention: none and only offers whose provider does not keep your content compete for the request.

Facts

Model name: gpt-oss-20b
Maker: OpenAI
Context window: 32.8K – 131.1K tokens
Input price: $0.010 to $0.037 per million tokens
Output price: $0.052 to $0.175 per million tokens
Providers: 3 independent providers
Regions: eu-central, us, us-west

Providers

Greenference: eu-central, $0.010 input and $0.052 output per million tokens, $0.0052 cache read per million tokens, 32.8K tokens of context
akash: us-west, $0.020 input and $0.100 output per million tokens, $0.020 cache read per million tokens, 131.1K tokens of context
Pera: us, $0.037 input and $0.175 output per million tokens, $0.037 cache read per million tokens, 122.9K tokens of context

Scores

HLE: 11.0, #153 of 273
GPQA: 68.8, #155 of 250
AA-LCR: 34.7, #190 of 237
SciCode: 38.9, #99 of 123
Terminal-Bench Hard: 10.6, #119 of 166
LiveCodeBench: 77.7, #21 of 130
AIME 2025: 89.3, #13 of 100
MMLU-Pro: 74.8, #73 of 133
τ²-Bench: 60.2, #94 of 169
IFBench: 65.1, #56 of 178

Scores: Artificial Analysis (https://artificialanalysis.ai/)

Request

curl https://router.inference.ai.exchange/v1/chat/completions \
  -H "Authorization: Bearer $LIQUID_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-oss-20b",
    "messages": [{"role":"user","content":"Say hello."}]
  }'

Questions

What is GPT: OSS 20B?: GPT: OSS 20B is a model from OpenAI.
How much does GPT: OSS 20B cost?: Offers now ask $0.010 to $0.037 per million input tokens and $0.052 to $0.175 per million output tokens. A job pays for the tokens it uses. The exchange picks the cheapest eligible offer.
What is the context length of GPT: OSS 20B?: The largest offer takes 131,072 input tokens. One job can generate up to 131,072 tokens. A smaller offer can take less. The exchange skips an offer that cannot take your message.
What can GPT: OSS 20B do?: Its providers list Reasoning effort, JSON object, JSON schema, Frequency penalty, Min-p sampling, Presence penalty, Repetition penalty, Seed, Stop sequences, Top-k sampling, Top-p sampling, Text, Tools.
What inputs and outputs does GPT: OSS 20B support?: Text → Text.
Which providers serve GPT: OSS 20B?: 3 independent providers quote it now, with 3 offers. The table above ranks them by the cost of a typical job.
How do I call GPT: OSS 20B?: Send "gpt-oss-20b" as the model in an OpenAI or Anthropic request. The API section links to setup steps for each client.
When was GPT: OSS 20B listed?: The exchange first listed it on 9/29/2026. This is not its release date.