llm-benchmarks
githubdrose.io

Claude Haiku Latest Benchmarks


Measured on bedrock

3 runs
Avg Tokens / Second
38.80
Decode Tokens / Second
Avg Time to First Token (ms)
890.00
Runs Analysed
3
Last Updated
Sep 13, 2026, 03:02 AM

Throughput distribution

Throughput over time

Claude Haiku Latestbedrock39tok/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 42 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
bedrockdirect38.8038.1039.90890.003

Frequently Asked Questions

How fast is Claude Haiku Latest?
The latest rolling average throughput is 38.80 tokens per second with an average time to first token of 890.00 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 13, 2026, 03:02 AM.

Related models