llama-3.2-3b by deepinfra Benchmarks

Stale Data

This model has not returned successful benchmark results recently. The data shown may be outdated.

Last benchmark: March 8, 2026

Benchmark Overview

Avg Tokens / Second

42.50

Avg Time to First Token (ms)

790.00

Runs Analysed

238

Last Updated

Mar 8, 2026, 06:05 AM

llama-3.2-3b streams at 42.50 tokens/second on average across the last 238 benchmark runs.
Performance fluctuated by 96.47 tokens/second (227.0% coefficient of variation), indicating variable behavior across benchmark runs.
Average time to first token is 790.00 ms (good latency), suitable for latency-sensitive workloads.
Latest measurements completed on Mar 8, 2026, 06:05 AM based on 238 total samples.

Distribution of throughput measurements showing performance consistency across benchmark runs.

Historical performance trends showing how throughput has changed over the benchmarking period.

Provider	Model	Avg Toks/Sec	Min	Max	Avg TTF (ms)
deepinfra	llama-3.2-3b	42.50	1.03	97.50	790.00

The latest rolling average throughput is 42.50 tokens per second with an average time to first token of 790.00 ms across 238 recent runs.

Benchmarks refresh automatically whenever the monitoring cron runs. The most recent run completed on Mar 8, 2026, 06:05 AM.