llm-benchmarks
githubdrose.io

Claude Opus 4.7 Benchmarks


Measured on anthropic

3 runs
anthropic site
Avg Tokens / Second
20.60
Decode Tokens / Second
Avg Time to First Token (ms)
1370.00
Runs Analysed
3
Last Updated
Sep 15, 2026, 12:01 PM

Throughput distribution

Throughput over time

Claude Opus 4.7anthropic21tok/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 28 tok/s (99th percentile) · dashed rule is the model's own mean · 1 cell carry fewer than 4 samples

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
anthropicvia OpenRouter20.6016.3026.201370.003

Frequently Asked Questions

How fast is Claude Opus 4.7?
The latest rolling average throughput is 20.60 tokens per second with an average time to first token of 1370.00 ms across 3 recent runs.
How often are these benchmarks updated?
Benchmarks refresh automatically whenever the monitoring cron runs. The most recent run completed on Sep 15, 2026, 12:01 PM.