Benchmark Overview
67.80
0.52
993
Dec 25, 2025, 11:49 PM
Key Insights
claude-3-haiku streams at 67.80 tokens/second on average across the last 993 benchmark runs.
Performance fluctuated by 45.10 tokens/second (66.5% coefficient of variation), indicating variable behavior across benchmark runs.
Average time to first token is 0.52 ms (excellent latency), suitable for latency-sensitive workloads.
Latest measurements completed on Dec 25, 2025, 11:49 PM based on 993 total samples.
Performance Distribution
Distribution of throughput measurements showing performance consistency across benchmark runs.
Performance Over Time
Historical performance trends showing how throughput has changed over the benchmarking period.
claude-3-haiku
Benchmark Samples
| Provider | Model | Avg Toks/Sec | Min | Max | Avg TTF (ms) |
|---|---|---|---|---|---|
| claude-3-haiku | 67.80 | 35.30 | 80.40 | 0.52 |
Frequently Asked Questions
The latest rolling average throughput is 67.80 tokens per second with an average time to first token of 0.52 ms across 993 recent runs.
Benchmarks refresh automatically whenever the monitoring cron runs. The most recent run completed on Dec 25, 2025, 11:49 PM.