Jungle Grid Pricing

Plans for every stage of AI infrastructure.

Platform plans are coming soon. Compute remains usage-based and is billed separately.

Jungle Grid platform plans

Available now

Developer

$0/ month

For individual developers exploring Jungle Grid and running early workloads.

Free managed inferencejg/text/fast-instruct

Run supported Jungle Grid managed models without selecting GPUs, managing containers, or supplying provider infrastructure.

Rex — Diagnose

Diagnoses supported errors originating in your workload and explains what needs attention.

  • Access to supported Jungle Grid managed models
  • 10 qualifying execution jobs free
  • Usage-based compute for other workloads
  • Inference, batch, training, and fine-tuning where supported
  • API, MCP, and CLI access
  • Pre-run estimates and intent-based routing
  • Infrastructure retries and failover
  • Logs and artifacts
  • Community support
Create account

No credit card required.

Coming Soon

Pro

$99/ month

For developers and small AI teams running production workloads.

Rex Pro — Diagnose + Fix

Planned supported fixes, workload retry, and recovery validation with up to 50 recoveries per month.

  • Everything in Developer
  • Rex Proplanned
  • Diagnose and fix supported workload errorsplanned
  • Automatic retry after Rex recoveryplanned
  • Recovery validationplanned
  • Up to 50 Rex Pro recoveries / monthplanned
  • Higher execution limitsplanned
  • Higher concurrencyplanned
  • Extended execution history and log retentionplanned
  • Advanced workload controlsplanned
  • Priority supportplanned
Coming SoonJoin Founding Deployment
Coming SoonMost Popular

Team

$299/ month

For AI companies running production workloads across a team.

Rex Pro — Automated Recovery + Team Controls

Planned workspace recovery with approvals and up to 200 Rex Pro recoveries per month.

  • Everything in Pro
  • Rex Pro across the workspaceplanned
  • Up to 200 Rex Pro recoveries / monthplanned
  • Rex approval controls where supportedplanned
  • Team collaborationplanned
  • Shared execution historyplanned
  • Multiple projects and environments where supportedplanned
  • Higher workload concurrencyplanned
  • Longer log and artifact retentionplanned
  • Team permissionsplanned
  • Priority technical supportplanned
  • Usage and spending controls where supportedplanned
  • Early capability accessplanned
Coming SoonJoin Founding Deployment
Talk to us

Enterprise

Custom

For organizations that need high-volume execution, security controls, dedicated support, or custom infrastructure requirements.

Rex — Custom Automation & Policies

Custom allowances, automation policies, and approval controls scoped to supported workload failures.

  • Everything in Teamplanned
  • Custom Rex limits and policiescustom
  • Custom execution limitscustom
  • Dedicated supportcustom
  • Architecture supportcustom
  • Custom provider policiescustom
  • Security and governance controlscustom
  • SLA, dedicated capacity, and private networkingplanned
Talk to the Founders

Included across Jungle Grid

The execution layer beneath every plan.

  • AI workload execution
  • Multi-provider routing
  • Pre-run estimates
  • Infrastructure retries and failover
  • Logs and lifecycle events
  • Artifacts
  • API / MCP / CLI access
  • Usage-based compute

Feature comparison

Compare Jungle Grid plans.

“Planned” identifies capabilities that are not represented as generally available today. Enterprise scope is confirmed directly with the Jungle Grid team.

Jungle Grid planned platform plan comparison
FeatureDeveloperProTeamEnterprise
Execution
Free managed inferenceAvailableAvailableAvailableAvailable
jg/text/fast-instructAvailableAvailableAvailableAvailable
10 qualifying execution jobs freeAvailable promotionAvailable promotionAvailable promotionAvailable promotion
Workload typesAvailableAvailableAvailableCustom scale
Usage-based computeAvailableAvailableAvailableAvailable
Pre-run estimatesAvailableAvailableAvailableAvailable
Intent-based / multi-provider routingAvailableAvailableAvailableCustom policies
Infrastructure retry / failoverAvailableAvailableAvailableCustom policies
Execution limitsAvailableHigher (planned)Higher (planned)Custom
ConcurrencyAvailableHigher (planned)Higher (planned)Custom
Interfaces
APIAvailableAvailableAvailableAvailable
MCPAvailableAvailableAvailableAvailable
CLIAvailableAvailableAvailableAvailable
Rex
Rex diagnosticsAvailableAvailableAvailableCustom
Rex Pro—PlannedPlannedCustom
Automated supported fixes—PlannedPlannedCustom
Retry after repair—PlannedPlannedCustom
Recovery validation—PlannedPlannedCustom
Rex recovery limit—Up to 50/mo (planned)Up to 200/mo (planned)Custom
Approval controls——PlannedCustom
Observability
LogsAvailableExtended (planned)Extended (planned)Custom
ArtifactsAvailableAvailableAvailableCustom
Execution historyAvailableExtended (planned)Shared (planned)Custom
RetentionAvailableExtended (planned)Longer (planned)Custom
Collaboration
CollaborationIndividual—PlannedCustom
Projects / environments——PlannedCustom
Shared execution history——PlannedCustom
Team permissions——PlannedCustom
Usage / spending controls——PlannedCustom
Support
Community supportAvailableAvailableAvailableAvailable
Priority support—PlannedPlannedCustom
Architecture support——PlannedCustom
Dedicated support———Custom
Enterprise controls———Custom / planned
SLA / dedicated capacity / private networking———Planned

Execution and recovery

Your infrastructure can recover. Your workload should too.

Jungle Grid

Handles infrastructure-side execution reliability.

Capacity · Placement · Provider failover · Scheduling · Infrastructure retries

Rex

Handles supported errors originating inside the workload.

Code · Dependencies · Runtime configuration · Model configuration · Validation

Jungle Grid keeps execution reliable. Rex finds and fixes errors in your workload.

Usage-based compute

Platform plan plus the compute your workload actually uses.

Jungle Grid compute remains usage-based and is separate from the platform subscription. Cost varies with workload requirements, runtime, and the capacity matched through multi-provider routing; inference, batch processing, fine-tuning, and full training have different compute profiles.

Estimate your AI workload cost →How Jungle Grid routes AI workloads →

FAQ

Frequently asked pricing questions.

Are these plans available today?

Developer access and current Jungle Grid execution capabilities are available today. Pro and Team platform subscriptions are coming soon.

Is compute included in the monthly plan?

No. Compute is usage-based and billed separately from the platform subscription.

What is Rex?

Rex is Jungle Grid's workload intelligence agent. It helps diagnose supported errors originating inside a customer's workload.

What is Rex Pro?

Rex Pro is planned to diagnose supported user-side workload failures and, where supported, apply fixes, retry the workload, and validate recovery.

Does Rex fix Jungle Grid infrastructure failures?

No. Jungle Grid handles infrastructure-side execution concerns such as capacity, placement, provider issues, retries, and failover. Rex focuses on supported problems originating inside the user's workload.

What counts as a Rex recovery?

A Rex recovery occurs when Rex handles a supported workload-originated failure through diagnosis and an automated repair and retry flow. Infrastructure retries and provider failures do not count as Rex recoveries.

Can I estimate compute before subscribing?

Yes. The workload estimator is available before you subscribe or create an account. Estimate your AI workload cost →

What if my company needs production support now?

Apply to the Founding Deployment Program or talk directly with the Jungle Grid founding team. Explore the Founding Deployment Program →

Deploy now

Want to deploy before the new plans launch?

Jungle Grid is working directly with a limited number of AI companies through the Founding Deployment Program.

Explore Founding DeploymentEstimate compute