delphi 1e21 3.4Bparams 46.3Btokens base wc386k lr1e5 sft

Open WeightModel weights are publicly available — can be downloaded and self-hosted

Released June 20264K context≈ 3,072 words

A compact base model from LAION trained on a relatively modest token budget of 46.3 billion tokens with supervised fine-tuning applied. At 3.4 billion parameters, it sits in the small-but-capable tier, where efficiency matters more than raw power. The SFT stage suggests it has been shaped for instruction-following behavior, though its 4096-token context window keeps it suited for shorter, focused interactions.

Capabilities

Capability scores are AI-generated based on model documentation, benchmarks, and technical specifications. Learn more

Factual Knowledge

Moderate

Instruction Following

delphi 1e21 3.4Bparams 46.3Btokens base wc386k lr1e5 sft

Open WeightModel weights are publicly available — can be downloaded and self-hosted

Released June 20264K context≈ 3,072 words

A compact base model from LAION trained on a relatively modest token budget of 46.3 billion tokens with supervised fine-tuning applied. At 3.4 billion parameters, it sits in the small-but-capable tier, where efficiency matters more than raw power. The SFT stage suggests it has been shaped for instruction-following behavior, though its 4096-token context window keeps it suited for shorter, focused interactions.

Capabilities

Capability scores are AI-generated based on model documentation, benchmarks, and technical specifications. Learn more

Factual Knowledge

Moderate

Instruction Following

Glossary

Base ModelA pretrained model that completes text patterns but hasn't been trained to follow instructions, serving as a starting point for customization through fine-tuning.Context WindowThe maximum number of tokens a model can process in a single conversation or prompt.Fine-TuningThe process of further training a pre-trained model on new data to adapt it for specific tasks or domains.Instruction-FollowingThe ability of a model to understand and execute specific tasks or commands given in natural language prompts.ParametersThe learned numerical values in a model — more parameters generally means more capacity but higher compute cost.Supervised Fine-tuningTraining a model on labeled examples to adapt it for a specific task or domain.Supervised Fine-tuningAdapting a pre-trained model using labeled examples where the correct outputs are known.TokenA small unit of text (a word, subword, or punctuation mark) that a language model breaks input into for processing.Token BudgetThe maximum number of tokens available to include retrieved context in a language model prompt.TokensThe basic units of text that a language model processes, typically representing words or word fragments.

Capabilities

Capabilities

Use Case Fit

Glossary