mercury
MoE · 284B total / 13B active
Fast and efficient. Built for volume.
High-volume work at low cost: drafting, transformation, classification and the many small calls a production system makes every day.
Models
Seven models. One API.
Every Based AI™ model is specialized on curated, expert-authored data and released with its architecture, parameter scale, intended use and evaluation record. Disclosure is standard policy for every model on the roster.
venus
MoE · 1.6T total / 49B active
Deep reasoning for demanding work.
Longer reasoning chains and harder synthesis. When a task outgrows mercury, venus takes it on.
mars
MoE · 744B total / 40B active
General-purpose and reliable.
Balanced capability across analysis, code and writing. Steady performance under sustained load.
saturn
Dense · 27B
Long tasks and large documents.
Sustained generation across long inputs. Consistent output that stays on task from start to finish.
logos
MoE · 2.8T total / 104B active
Language and long-form reading.
Built for written work: long documents, close reading and prose that holds up to editing.
grug
Dense · 27B
Short answers. Direct and plain.
Ask a question and receive the answer, stated plainly and briefly.
based
MoE · 744B total / 40B active
The general default.
Our first choice for most work. Broad capability with a direct, sourced and measured style.
Reserved for your custom model
Finetuning on the platformDisclosure policy.
Four rules apply to every model on the roster, including custom models built for clients.
01
Every model publishes its architecture and parameter scale. If either changes, the model receives a new name.
02
Every model states its intended use and its known limits.
03
Every model documents the provenance of the data used to specialize it.
04
Every model is evaluated against task-specific harnesses before it joins the roster. Retired models stay listed and are marked retired.
Architecture.
The roster uses two transformer architectures. Each model lists its architecture and parameter scale, so you can judge capacity, latency and cost before you choose.
Mixture-of-Experts
mercury · venus · mars · logos · based
A router activates a small share of the parameters for each token. This gives large total capacity at a lower serving cost, suited to broad knowledge, reasoning and long-context work.
Dense
saturn · grug
Every parameter is active for every token. Latency and memory use are predictable, and the model is simpler to specialize and deploy. Suited to sustained generation and focused tasks.
V — Access
Build with Based AI.
API access, custom builds and finetunes start with a short request. Tell us what you are building and a member of our team will reply.