Nox Vela 2
The fast model on every plan — now more accurate
Vela 2 is the everyday model of the Nox family and the one available on every plan, including Free. It builds on every previous Vela generation with higher accuracy at the same speed. It runs on a 30-billion-parameter hybrid mixture-of-experts engine that activates 3 billion parameters per token.
Vela 2 is the newest generation of Nox's fast models, built on every previous Vela with higher accuracy at the same speed. It answers with the 1.3 generation's cleaner organisation while keeping latency low enough that replies land in roughly the time the question takes to read.
The engine combines Mamba sequence layers with Transformer attention in a mixture-of-experts design: 30 billion total parameters with 3 billion active per token. It is a genuine reasoning model, with reasoning behavior configurable to match the task while keeping speed as the headline property.
Its context window is roughly 262,000 tokens, which keeps long conversations fully in view. Vela itself is text-only; when a photo is attached, Nox handles it automatically with a compatible vision model. It also carries the family's shared tools: inline charts, connected-app actions, and consent-gated live web access with cited sources.
Like every Nox model, Vela sits behind Leo, the deterministic safety layer whose measured recall and false-positive rate are published from a generated evaluation and identical family-wide — the same gate in front of the fastest model as the flagship.
Vela 2 is available on every plan, including Free, and is the active everyday model.
Capabilities
- Fast replies from a 3-billion-active-parameter hybrid Mamba-Transformer mixture-of-experts engine
- Configurable built-in reasoning behavior
- ~262,000-token context window
- Photo attachments handled automatically, plus inline charts and connected-app actions
- Consent-gated live web access with every source cited
Model details
- Architecture — Hybrid Mamba-Transformer mixture-of-experts — 30 billion parameters total, 3 billion active per token
- Context window — ~262,000 tokens
- Reasoning — Built-in reasoning with configurable behavior
- Latency — The fast everyday model
- Photo understanding — Handled automatically by Nox with a compatible vision model
- Live web access — Yes — consent-gated; Nox asks before searching, then cites every source it uses
- Safety layer — Runs behind Leo, Nox's deterministic red-flag detector, on every message
- Available on — Free, Pro, and MAX
Availability: Available on every plan, including Free.