28 models. One fleet. Served by Riven.
Every model below runs on Riven’s own inference fleet and is live on the API right now — no resold third-party lanes, no hidden markup. Each lane was individually probe-verified before it earned a place in this catalog.
Riven tier lanes keep a stable name and intent while Riven upgrades the backing engine behind the lane. Pay-as-you-go meters per token at each lane’s current rate.
The Riven tier family
Seven intent-based lanes cover the full speed/quality range — the same lanes shown in the model picker in the Riven apps.
Riven Auto
routed laneRoutes each request to the right lane. Billed at the lane it uses.
All plans · PAYG $4.09/1M output
Riven Spark
1xThe fastest lane — built for latency-sensitive traffic.
All plans · PAYG $2.73/1M output
Riven Bolt
1.5xEveryday chat, strong quality at low cost.
All plans
Riven Nova
3.75xThe balanced flagship. Recommended pinned lane for production.
All plans
Riven Atlas
6.7xDeep reasoning for hard, multi-step problems.
Paid plans · PAYG $10.90/1M output
Riven Research
2.3xLong-form research and report-grade work.
Paid plans · PAYG $6.82/1M output
Riven Apex
13.3xThe most capable lane, with the largest context window.
Paid plans · PAYG $16.36/1M output
Behind the curtain
Tier names are stable; the engines behind them are upgraded in place. Raw fleet ids stay in the catalog below.
The full catalog — 28 lanes, all riven-owned
Grouped by role: tier lanes, tier aliases, fine-tuned lanes, legacy compatibility IDs, fleet engines, and embeddings. Compatibility IDs are accepted for old integrations and rewritten to their canon lane at the gateway before serving.
Tier lane (7)
| Model ID | Lane / routes to | Availability | Input /1M | Output /1M |
|---|---|---|---|---|
| riven-best | Riven Auto | All plans | $1.36 | $4.09 |
| riven-instant | Riven Spark | All plans | $0.91 | $2.73 |
| riven-fast | Riven Bolt | All plans | — | — |
| riven-core | Riven Nova | All plans | — | — |
| riven-pro | Riven Atlas | Paid plans | $3.63 | $10.90 |
| riven-research | Riven Research | Paid plans | $2.27 | $6.82 |
| riven-max | Riven Apex | Paid plans | $5.45 | $16.36 |
Tier alias (4)
| Model ID | Lane / routes to | Availability | Input /1M | Output /1M |
|---|---|---|---|---|
| riven-atlas | riven-pro | Paid plans | $3.63 | $10.90 |
| riven-apex | riven-max | Paid plans | $5.45 | $16.36 |
| riven-bolt | riven-fast | All plans | — | — |
| riven-spark | riven-fast | All plans | — | — |
Fine-tuned lane (2)
| Model ID | Lane / routes to | Availability | Input /1M | Output /1M |
|---|---|---|---|---|
| riven-default | riven-default | All plans | — | — |
| riven-chat | Riven Chat | All plans | — | — |
Compatibility ID (14)
| Model ID | Lane / routes to | Availability | Input /1M | Output /1M |
|---|---|---|---|---|
| riven-chat-sm | riven-fast | All plans | — | — |
| riven-chat-md | riven-default | All plans | — | — |
| riven-chat-lg | riven-best | All plans | $1.36 | $4.09 |
| riven-chat-xl | riven-best | All plans | $1.36 | $4.09 |
| riven-tiny | riven-fast | All plans | — | — |
| riven-code | riven-best | Paid plans | $1.36 | $4.09 |
| riven-vision | riven-best | Paid plans | $1.36 | $4.09 |
| riven-agent | riven-best | All plans | $1.36 | $4.09 |
| riven-web | riven-best | All plans | $1.36 | $4.09 |
| riven-frontier | riven-best | All plans | $1.36 | $4.09 |
| rvn-assistant-v2 | riven-best | All plans | $1.36 | $4.09 |
| riven-deep | riven-max | Paid plans | $5.45 | $16.36 |
| riven-ultra | riven-max | All plans | $5.45 | $16.36 |
| riven-ultra-research | riven-max | Paid plans | $5.45 | $16.36 |
Embeddings (1)
| Model ID | Lane / routes to | Availability | Input /1M | Output /1M |
|---|---|---|---|---|
| nomic-embed-text | nomic-embed-text | All plans | — | — |
Served by Riven, not resold
Every lane runs on Riven’s own inference fleet. Tier names and intents are stable, the engines behind them are upgraded in place, and usage is metered per token at each lane’s current rate — derived from actual fleet cost, not a hidden margin on someone else’s price list.
See plans