GPT-OSS 20B on Groq
Lower-cost Groq-hosted GPT-OSS tier for reasoning, tool use, and structured outputs.
- Input
- $0.07 / 1M tokens
- Output
- $0.30 / 1M tokens
- Cached read
- $0.04 / 1M tokens
- Cached write
- $0.07 / 1M tokens
- Batch discount
- —%
- Source
- GPT-OSS 20B on Groq pricing
- Verified
- Apr 2, 2026 (High)
Capabilities
- Modalities
- text→text
- Capabilities
- reasoningwebSearchbatchSupportcodeExecutionpromptCachingfunctionCallingstructuredOutputs
- Strengths
- Cheap hosted reasoning, Very low latency
- Tradeoffs
- Lower ceiling than GPT-OSS 120B
Official Links
Benchmark Coverage
| Benchmark | Version | Score | Date | Source | Notes |
|---|
Release History
| Release | Alias | Lifecycle | Release Date | Deprecation | Shutdown | Summary |
|---|---|---|---|---|---|---|
| GPT-OSS 20B on Groq | groq-gpt-oss-20b | Active | Aug 1, 2025 | — | — | Current published model family snapshot. |
Host Coverage
| Host | Type | Context | Pricing Note | Differences |
|---|---|---|---|---|
| Groq API | first-party | 131.1K | Reference hosted GPT-OSS pricing on Groq. | Prompt caching; Structured outputs |
Migration Guidance
Budget Groq-hosted reasoning tier for workloads that do not need 120B depth.
Replacement models: groq-gpt-oss-120b
Change Events
| Date | Type | Title | Description | Source |
|---|---|---|---|---|
| Aug 1, 2025 | family_added | GPT-OSS 20B on Groq published | Initial public model family launch. | GPT-OSS 20B on Groq release notes |
Other models from Groq
GPT-OSS 120B on Groq, Groq Compound, Llama 3.1 8B Instant on Groq, Llama 3.3 70B Versatile on Groq, Llama 4 Scout on Groq, Qwen3 32B on Groq