SynapseX model reference
This page lists every SynapseX model you can pick in SynapseX Chat, with its exact id, context window, output limit and list price. Use it to choose the cheapest model that fits your task, and to copy the exact id string each surface expects.
Scope. This table covers the models verified as selectable in SynapseX
Chat. The Platform API serves its own catalog at
GET https://platform.synapsex.ai/api/v1/models, which is not the same list —
and not every id in it is callable by an API key. Do not treat the prices below
as the Platform API price list, and see Models on the API before
you pick an id for API use.
Which SynapseX models can I choose in Chat?
| Id | Name | Best for | Context | Max output | Input / 1M | Output / 1M |
|---|---|---|---|---|---|---|
synapsex:synapsex-atlas-7b-v1 | SynapseX Atlas | Compact, fast model for general tasks and high volume. Default | 128,000 | 16,000 | $0.10 | $0.40 |
synapsex:synapsex-forge-code-14b-v1 | SynapseX Forge | Code generation and review | 256,000 | 8,000 | $0.40 | $1.20 |
synapsex:synapsex-vault-enterprise-v1 | SynapseX Vault | Enterprise work needing long context and compliance | 131,000 | 16,000 | $0.35 | $1.40 |
synapsex:synapsex-nexus-v2 | SynapseX Nexus | Fast multi-agent orchestration across code, chat and tools | 131,000 | 8,000 | $0.10 | $0.40 |
synapsex:synapsex-oracle-v1 | SynapseX Oracle | Advanced reasoning for complex tasks (reasoning model) | 131,000 | 8,000 | $0.75 | $3.00 |
synapsex:synapsex-quasar-v1 | SynapseX Quasar | Quantum computing research and frontier reasoning | 200,000 | 32,000 | $3.50 | $14.00 |
Prices are US dollars per one million tokens. Context and max output are in tokens.
Atlas is the model a new chat starts on. To change it, press Cmd+Shift+P (or
Ctrl+Shift+P) in the composer — see
/reference/keyboard-shortcuts.
Which models do not count against the daily pro-model limit?
Signed-in accounts get 1,000 messages a day. Models outside the free list additionally consume a separate 500-per-day counter.
| Model | Consumes the 500/day pro-model counter? |
|---|---|
| SynapseX Atlas | No |
| SynapseX Forge | No |
| SynapseX Vault | Yes |
| SynapseX Nexus | Yes |
| SynapseX Oracle | Yes |
| SynapseX Quasar | Yes |
If you are working through a long session and do not need frontier reasoning, staying on Atlas or Forge keeps the pro-model counter untouched. Full limit behaviour: /chat/limits-and-errors.
How is a model id written on each surface?
The same model has three spellings. They are not interchangeable.
| Surface | Format | Example |
|---|---|---|
| SynapseX Chat | synapsex:<id> | synapsex:synapsex-atlas-7b-v1 |
SynapseX CLI (-m) | synapsex/<id> | synapsex/synapsex-atlas-7b-v1 |
| Platform API request body | the bare id | synapsex-atlas-7b-v1 |
The CLI carries its own catalog, which is not identical to the Chat list. Run
synapsex-code models to see exactly which ids your installed CLI accepts.
How do I verify this list myself?
The Chat catalog is public — no key required:
curl -s https://chat.synapsex.ai/api/modelsThat returns the six models above with their id, name, description,
contextWindow, inputCost and outputCost.
The Platform API catalog needs your key:
curl -sS https://platform.synapsex.ai/api/v1/models \
-H "Authorization: Bearer sk-synapsex-YOUR-API-KEY"Read that at runtime rather than hard-coding ids — but treat it as a catalog, not a permission list: an id can be listed there and still be refused with a 403 that names the ids you may use. Response shape and error codes: /reference/api-endpoints.