The small, fast LiquidAI model is LFM2.5-1.2B-Thinking (free) — the efficient tier at 106 tokens/sec and Free 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. LFM2.5-1.2B-Instruct (free) (66 t/s) and LFM2-24B-A2B (57 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 LiquidAI model is LFM2.5-1.2B-Thinking (free) — the efficient tier at 106 tokens/sec and Free 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. LFM2.5-1.2B-Instruct (free) (66 t/s) and LFM2-24B-A2B (57 t/s) round out the top three.
LFM2.5-1.2B-Instruct (free) (66 t/s) is the closest alternative on this metric, followed by LFM2-24B-A2B (57 t/s). See the full ranking above for the tradeoffs.
modelgrep tracks 3 LiquidAI models with live benchmarks, speed, latency and per-provider pricing, led on intelligence by LFM2-24B-A2B. 3 of them qualify for this ranking.