News
Frontier‑Cost LLM Classifier Benchmark Shows 1% Production Cost
Airbnb fine‑tuned a 2‑billion‑parameter model achieving 0.902 F1 with 35 ms median latency, costing roughly 1 % of frontier API pricing per GPU.
News
Airbnb fine‑tuned a 2‑billion‑parameter model achieving 0.902 F1 with 35 ms median latency, costing roughly 1 % of frontier API pricing per GPU.