Your sovereign, private GPT OSS 120B
GPT-OSS 120B is OpenAI's open-weight model (Apache 2.0) for powerful reasoning, agentic tasks and versatile developer use cases: 117B parameters with 5.1B active, MXFP4-quantized to fit a single 80GB GPU. It features configurable reasoning effort (low/medium/high), full chain-of-thought access and native function calling, browsing and Python execution. A reasoning model in the large tier, served on ToothFairyAI's AU, EU and US endpoints — private by default, billed in US dollars per million tokens at the published rates below. No training on your data, no vendor lock-in.
How GPT OSS 120B compares
Independent intelligence index for GPT OSS 120B against every peer in the Large tier (and the tier above where needed) — same source, same tests.
Latency — end-to-end seconds for a 500-token answer
Capability indexes
Domain-weighted agentic performance for GPT OSS 120B.
Capability index vs large peers
The evaluations behind the index
Independent benchmark results for GPT OSS 120B — Elo scores as published; pass rates shown as percentages.
Per-evaluation economics
| Evaluation | Score | Time / task | Output tokens / task |
|---|---|---|---|
| AA-Briefcase | 0.0% | 8.4 min | 97717 |
| GDPval-AA | 595.50 | 2.7 min | 31411 |
| AutomationBench-AA | 0.2% | 0.2 min | 2183 |
| Terminal-Bench 4.0 | 0.0% | 1.8 min | 20583 |
| SciCode | 34.0% | 0.5 min | 5707 |
| Humanity's Last Exam | 19.6% | 2.0 min | 22890 |
| GDP.pdf | 4.0% | 0.7 min | 7681 |
| CritPt | 1.1% | 2.5 min | 29812 |
| AA-Omniscience | -4925.0% | 0.2 min | 1916 |
| AA-LCR (long context) | 52.0% | 0.2 min | 1831 |
Run GPT OSS 120B privately
One key, one endpoint, three jurisdictions.
- Size tierLarge
- Serving routeStandard
- Input rate$0.84/1M
- Output rate$2.10/1M
- Cached input rate—
- Context window131K tokens
- Max output66K tokens
- ReasoningYes — emits thinking traces
- Vision & video inputText only
- Tool callingSupported
- Data residencyAU, EU and US endpoints — processing stays inside the region you call