OpenAI · live price
gpt-oss-20b API 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...
Small and cheap (under $1 per 1M)Tool callingReasoning
- Input
- $0.075
- per 1M tokens
- Output
- $0.3
- per 1M tokens
- Cached input
- $0.037
- per 1M tokens
- Context
- 131K
- 66K max output
List price from OpenRouter, which passes OpenAI's price through without markup. Refreshed hourly. Released August 5, 2025. Model id openai/gpt-oss-20b.
What gpt-oss-20b costs in practice
At list price, before any batch or caching discount.
1,000 chat messages
$0.23
1,000 input + 500 output tokens each
1,000 RAG questions
$0.75
8,000 input + 500 output tokens each
100 coding-agent steps
$0.29
30,000 input + 2,000 output tokens each, no caching
gpt-oss-20b price history
Price last changed on 2026-11-21. We record prices daily since 2026-10-01.
gpt-oss-20b price by host
Darkbloom serves it below list price right now. Route to it to pay less on every token.
- Darkbloom$0.017 in · $0.085 out · 131K ctx · 100.0% uptime 24h
- AkashML$0.018 in · $0.09 out · 131K ctx · 99.8% uptime 24h
- DeepInfra$0.024 in · $0.112 out · 131K ctx · 0.0% uptime 24h
- CoreWeave$0.03 in · $0.13 out · 131K ctx · 97.0% uptime 24h
- Parasail$0.03 in · $0.15 out · 131K ctx · 99.9% uptime 24h
- SiliconFlow$0.04 in · $0.18 out · 131K ctx · 95.4% uptime 24h
- Amazon Bedrock$0.07 in · $0.15 out · 131K ctx · 100.0% uptime 24h
- Groq$0.075 in · $0.3 out · 131K ctx · 99.5% uptime 24h
- PrimeIntellect$0.2 in · $0.3 out · 131K ctx · 95.8% uptime 24h
Where to buy OpenAI credits for less
gpt-oss-20b costs the same per token everywhere it's sold through OpenRouter. What changes is the price of the credits. Live quotes for $100 of usage:
gpt-oss-20b cost calculator
Enter your monthly tokens to price your own usage.
Usage at list price$6.75/ month(quotes below are for $7.00: purchases run from $5.00 to $10,000.00)
Cheaper alternatives to gpt-oss-20b
Other labs' models in the same price range or the one below, cheaper per token.
| Model | Input / 1M | Output / 1M |
|---|---|---|
| DeepSeek: DeepSeek V4.1 Flash | $0.05 | $0.199 |
| DeepSeek: DeepSeek V4 Flash 0731 | $0.044 | $0.3 |
| Qwen: Qwen3.7 Flash | $0.03 | $0.13 |
| DeepSeek: DeepSeek V4 Flash 0423 | $0.09 | $0.18 |
| Google: Gemma 4 26B A4B | $0.068 | $0.225 |
| Qwen: Qwen3.5-Flash | $0.065 | $0.26 |
Other OpenAI models
| Model | Input / 1M | Output / 1M |
|---|---|---|
| gpt-oss-120b | $0.09 | $0.36 |
| gpt-oss-safeguard-20b | $0.075 | $0.3 |
| GPT-5.1-Codex-Mini | $0.25 | $2 |
| GPT-5.1-Codex | $1.25 | $10 |
| GPT-5.1 | $1.25 | $10 |
| GPT-4.1 | $2 | $8 |