Join the beta

OpenAI · 2025-08-05

gpt-oss-20b pricing

gpt-oss-20b is an open-weight 21B parameter model released by OpenAI under the Apache 2.0 license.

Prices from OpenRouter’s models API, updated October 29, 2026

Input$0.018per 1M tokens
Output$0.09per 1M tokens
Cached input$0.00950% off input
Context131Ktools · reasoning

What gpt-oss-20b costs a month

WorkloadStandardBatch API
Support chatbot: 1,000 chats a day, 1,500 tokens in, 400 out$1.89$2.42
Daily summaries: 10,000 a day, 3,000 tokens in, 300 out$24.30$31.68
Coding agent: 400 calls a day, 60,000 tokens in (80% cached), 2,000 out$9.94$11.33

30 days. Cached input billed at the cache-read price. Use your own numbers.

Other GPT models

ModelInput /1MOutput /1MCache read /1MContextChatbot, per month*
GPT-6.1 Terra$1$5$0.051.05M$105
GPT-6.1 Terra Pro$1$5$0.051.05M$105
GPT-6.1 Sol$2$10$0.101.05M$210
GPT-6.1 Sol Pro$2$10$0.101.05M$210
GPT-6 Luna$0.10$0.50$0.011.05M$10.50
GPT-6 Luna Pro$0.10$0.50$0.011.05M$10.50

* Support chatbot: 1,000 chats a day, 1,500 tokens in, 400 out, 30 days, no caching. Prices in US dollars per 1M tokens.

Related

gpt-oss-20b: questions

How much does gpt-oss-20b cost?

$0.018 per 1M input tokens and $0.09 per 1M output tokens, with cached input at $0.009, or $0.024 / $0.112 through the batch API. A 1,000-token prompt with a 500-token answer costs about $0.0001.

What is gpt-oss-20b’s context window?

131,072 tokens, with up to 32,768 output tokens per response.

Know what each model really costs you.

Connect OpenRouter and Glow shows spend by model, key and product, and flags the swap that saves the most at the same quality.

One email when your invite is ready. No spam, unsubscribe any time.

  • Read-only keys
  • Never shown again
  • 2FA + Face ID
  • 7-day trial