llm-benchmarks
githubdrose.io

gpt-oss-120b Benchmarks


Measured on groq

13 runs
groq site
Avg Tokens / Second
72.10
Decode Tokens / Second
465.00
Avg Time to First Token (ms)
790.00
Runs Analysed
13
Last Updated
Sep 4, 2026, 03:11 PM

Throughput distribution

Throughput over time

gpt-oss-120bgroq179tok/s

Generated throughput, including reasoning tokens — so a thinking model's line does not collapse; the table ranks on visible tokens · one cell per model, showing its best-covered provider · models with at least 4 samples in the window first, fastest of those at the top · shared vertical scale, 0 to 270 tok/s (99th percentile) · dashed rule is the model's own mean

Every lane measuring this model

1 measured · live first · a lane is a provider, a catalogue id and a transport
ProviderLaneAvg Toks/SecMinMaxAvg TTF (ms)n
groqvia OpenRouter72.1040.1097.80790.0013

Frequently Asked Questions

How fast is gpt-oss-120b?
The latest rolling average throughput is 72.10 tokens per second with an average time to first token of 790.00 ms across 13 recent runs. Once tokens are flowing the decoder delivers about 465.00 tokens per second, measured from the gaps between stream chunks with time to first token and reasoning excluded.
How often are these benchmarks updated?
Benchmarks refresh automatically whenever the monitoring cron runs. The most recent run completed on Sep 4, 2026, 03:11 PM.

Related models