Get started
Models
The model picker lets you choose which foundation model drives a new Nomic agent session. By default, sessions use Auto, which lets Nomic pick the current default model in the backend for that session. All prices below are passed through at the provider's API rate and are quoted per 1M tokens. Provider rates are the same on every plan.
Picking a model
- Auto — Recommended default. Nomic picks the current model and uses Medium effort.
- Sonnet 5.5 — Best balance of speed, cost, and quality for daily Nomic agent use and most workflows.
- Sonnet 5.5 Extra High — Same model, Extra High effort. For exhaustive work — for example a comprehensive drawing review.
- Opus 5.5 — Anthropic's newest Opus model. Use for the most difficult tasks.
- Opus 4.8 — Use for the hardest reasoning, ambiguous specs, and high-stakes reviews. About 2.5× Sonnet 5.5's per-token price.
- Haiku 4.5 — Use for quick lookups, exploration, and high-volume automation. About half Sonnet 5.5's per-token price.
More models (turn Auto off, then expand):
- Sonnet 4.6 — The previous Sonnet generation.
- Sonnet 5 — More intelligent than Sonnet 4.6 with higher token usage.
- Opus 4.6 — Use for demanding reasoning when you want Opus behavior with the previous Opus generation. About 2.5× Sonnet 5.5's per-token price.
Two Google models are available to organizations that have them enabled. If your picker shows only Auto and the Anthropic models, these are not turned on for you:
- Gemini 3.5 Flash — Use for fast Google-based agentic tasks and very long context (up to 1M tokens).
- Gemini 3.1 Pro — Under More models. Google's Pro model for strong reasoning and tool use, or when your team prefers a Google-based inference path.
The closed list is one item per mean job, not every model × effort combination. On Claude, set Low, Medium, High, or Extra High on the selected model (Medium is unlabeled). Same per-token price; higher effort uses more tokens.
The model is fixed for the session — start a new session to change it. You can still switch effort on the same model. Tools, citations, and file context stay the same.
Some helper steps use a fixed model regardless of the picker: quick file exploration runs on Haiku 4.5, locating geometry on drawings uses Gemini 3.5 Flash, and drawing-judgment checks use Gemini 3.1 Pro. These are billed at those models' rates. The Workflow Builder's assistant opens on Opus 5.5 at Medium effort.
Effort
Low — less reasoning, faster, cheaper. Medium — deeper reasoning before acting. High — thorough reasoning. Extra High — longest reasoning and largest tool budget; costs the most.
Pricing
| Model | Provider | Input | Output | Cache write | Cache read |
|---|---|---|---|---|---|
| Haiku 4.5 | Anthropic | $1.00 | $5.00 | $1.25 | $0.10 |
| Sonnet 5.5 | Anthropic | $2.00 | $10.00 | $2.50 | $0.20 |
| Sonnet 4.6 | Anthropic | $3.00 | $15.00 | $3.75 | $0.30 |
| Sonnet 5 | Anthropic | $3.00 | $15.00 | $3.75 | $0.30 |
| Opus 5.5 | Anthropic | $4.00 | $20.00 | $5.00 | $0.20 |
| Opus 4.8 | Anthropic | $5.00 | $25.00 | $6.25 | $0.50 |
| Opus 4.6 | Anthropic | $5.00 | $25.00 | $6.25 | $0.50 |
| Gemini 3.5 Flash | $1.50 | $9.00 | — | $0.15 | |
| Gemini 3.1 Pro | $2.00 | $12.00 | — | $0.20 |
Prices are USD per 1M tokens. Anthropic models charge separately for cache writes; Gemini does not. On later turns in the same session, cache reads are typically 10× cheaper than fresh input — the Nomic agent takes advantage of this automatically as the conversation continues.
Nomic Token Fee
In addition to the provider rates above, Nomic charges a Token Fee of $0.50 per million tokens. The fee is the same for every token type (input, output, cache write, and cache read) and does not vary by model. This fee supports our agent orchestration optimizations and model infrastructure improvements when delivering our agents.