LLM model comparison
unchecked params use each model's defaults (Anthropic always needs max tokens — defaults to 4096)
Closed source idle
latency
—
1st token
—
tokens
—
Closed-model response appears here.
VS
Open source idle
latency
—
1st token
—
tokens
—
Open-model response appears here.