gpt-oss (OpenAI) Pricing
All plans and pricing as of 2026-04-17
Self-hosted (Free, Apache 2.0)
- ✓First OpenAI open-weight release ever
- ✓Weights on Hugging Face + Ollama + llama.cpp + vLLM
- ✓Apache 2.0 license -- unrestricted commercial use
- ✓No telemetry, no phone-home, runs fully offline
API (OpenRouter / Together / Fireworks)
- ✓gpt-oss-120b: ~$0.15 in / $0.60 out
- ✓gpt-oss-20b: ~$0.07 in / $0.30 out
- ✓Competitive per-token pricing across hosted providers
Is gpt-oss (OpenAI) Worth the Price?
Value Score: 10/10
Overall Score: 8.1/10 · Developers who want OpenAI-brand open-weight reasoning models for self-hosting or fine-tuning. Particularly good for single-GPU deployments (gpt-oss-120b on one 80GB card) or edge-device reasoning (gpt-oss-20b on 16GB consumer GPUs / Apple Silicon). Also good as a reliable baseline when comparing newer open-weight releases.
gpt-oss remains historically important as the first OpenAI open-weight release (August 2025), and the 120b model on a single 80GB GPU is still one of the cleanest single-card frontier-reasoning options in the open-weight category. By April 2026 it is no longer the bleeding edge -- DeepSeek V3.2, GLM-5.1, and Qwen 3.6 have all shipped stronger models -- but gpt-oss's combination of OpenAI brand + genuine Apache 2.0 + single-GPU 120b sizing makes it a durable default in any open-weight evaluation matrix. Worth adding to any shortlist; probably not first pick unless the OpenAI brand association matters for your stack.
How gpt-oss (OpenAI) Pricing Compares
| Tool | Free Tier | Starting Price | Value Score | Overall |
|---|---|---|---|---|
| gpt-oss (OpenAI)(this tool) | Yes | $0 | 10/10 | 8.1 |
| Qwen (Alibaba) | Yes | $0 | 10/10 | 8.8 |
| MiniMax M2 / M2.5 | Yes | $0 | 9.5/10 | 8.4 |
| Gemma 4 (Google) | Yes | $0 | 10/10 | 8.3 |
| IBM Granite 4.0 | Yes | $0 | 9.5/10 | 8.2 |
| Kimi K2.5 (Moonshot) | Yes | $0 | 8.5/10 | 8.1 |