H2O-Danube2-1.8B
H2O's second-gen 1.8B small model trained on 3T tokens.
About
H2O-Danube2-1.8B is H2O.ai's second-generation small language model, announced in July 2024 as an incremental upgrade over the original Danube 1.8B. It was trained on an additional 2 trillion tokens on top of the first generation's 1 trillion, keeping the 1.8B parameter count and 8K context while improving quality.
Released under Apache 2.0 in base, SFT, and chat-tuned versions, it targets offline and mobile use cases such as summarization, RAG, and on-device chat. The chat version was tuned with supervised fine-tuning plus DPO.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 7/10
Scored 7/10 (metrics tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 126; Trending score 72/100 from researcher assessment.