The small, fast Anthropic model is Claude Haiku 4.5 — the efficient tier at 82 tokens/sec and $1.00 per million input tokens. It trades a few points of raw intelligence for speed and cost, the right call for high-volume, latency-sensitive work. Claude 3 Haiku (68 t/s) and Claude 3.5 Haiku (35 t/s) round out the top three.
Compact, efficient models — the small/mini/flash/haiku tier — ranked by output speed. These trade a little raw intelligence for low cost and high throughput, which is the right tradeoff for chat, classification, extraction and other high-volume work.
The small, fast Anthropic model is Claude Haiku 4.5 — the efficient tier at 82 tokens/sec and $1.00 per million input tokens. It trades a few points of raw intelligence for speed and cost, the right call for high-volume, latency-sensitive work. Claude 3 Haiku (68 t/s) and Claude 3.5 Haiku (35 t/s) round out the top three.
Claude 3 Haiku (68 t/s) is the closest alternative on this metric, followed by Claude 3.5 Haiku (35 t/s). See the full ranking above for the tradeoffs.
modelgrep tracks 16 Anthropic models with live benchmarks, speed, latency and per-provider pricing, led on intelligence by Claude Fable 5. 3 of them qualify for this ranking.