Marketplace

C3 aggregates GPU and CPU capacity from multiple data centers. When you submit a job, we find available compute at competitive rates, with no need to manage cloud accounts or hunt for capacity yourself.

Available hardware

Hardware	Memory	Best for
L40 class	Current range from c3 list	Recommended for new jobs, training, inference, and general-purpose GPU workloads
A100 class	Current range from c3 list	Larger memory and higher-throughput CUDA workloads
H100 class	Current range from c3 list	Highest-end accelerator workloads when H100 capacity is enabled
cpu-d3-4vcpu-16gb	16 GB RAM	Small CPU-only jobs, package builds, and smoke tests
cpu-d3-96vcpu-384gb	384 GB RAM	Large AMD CPU scientific computing, simulation, compilation, and data processing
cpu-e2-4vcpu-16gb	16 GB RAM	Small Intel CPU-only jobs, package builds, and smoke tests
cpu-e2-48vcpu-192gb	192 GB RAM	Large Intel CPU scientific computing, simulation, compilation, and data processing
cpu-n1-4vcpu-4gb	4 GB RAM	Small NextGen CPU-only jobs and smoke tests
cpu-n1-16vcpu-32gb	32 GB RAM	Larger NextGen CPU-only jobs and data processing

Omit hardware in your .c3 config, or set hardware: l40, hardware: a100, or hardware: h100, to let C3 choose within that primary class. gpu remains a compatibility alias for GPU profiles. The job records the concrete C3 subclass/profile it lands on, such as l40s-d-32x192, a100-80gb-pcie, or h100-80gb, and bills the selected provider/profile rate. Run c3 list or open the live hardware status page for the public hardware view with current pricing, memory ranges, and C3-level availability tiers. Run c3 list -al to see the concrete C3 subclass/profile codes inside each class, or c3 list --class h100 --all for one class. Expanded rows identify the provider and show that provider's current C3 billed price; use c3 list --provider crusoe --all to filter them. Provider-native SKU names, regions, exact stock counts, warm capacity, and internal readiness reasons are not exposed before a job is matched. A4000 is retired from C3 public stock.

CPU profiles are exact C3 hardware profiles. To request CPU capacity, set hardware: to one of the CPU profiles shown by c3 list, such as cpu-d3-96vcpu-384gb or cpu-e2-48vcpu-192gb; do not set gpu:. C3 infers internally that the selected profile is CPU hardware with no CUDA accelerator, so CPU jobs run without NVIDIA drivers or Docker GPU flags. For Docker images that are intended for CPU-only execution, set docker.requires_accelerator: none; CUDA images should set docker.requires_accelerator: cuda so C3 rejects an incompatible CPU route before your script starts.

cpu itself is a display group, not a selectable profile. Submitting hardware: cpu returns a validation error that points to c3 list --class cpu --all.

All CPU availability is currently experimental. CPU jobs remain usable and billable, but capacity and C3-managed images are best-effort and may change. c3 list, the public status page, and deploy route preview show this maturity stage separately from the current high/medium/low availability tier. Deploying to an experimental CPU route prints a warning but does not block submission.

Pricing

Compute

You're billed per second of actual compute time—not for time spent queuing or provisioning. New accounts receive £10 in free credits. Check your balance:

c3 balance

Plans

Subscriptions control maximum active compute concurrency, denominated in chips. One submitted machine is one chip, whether the job runs on GPU or CPU hardware. Compute is still billed per second, and storage usage is tracked separately for usage-based billing. Every new account starts with £10 in compute credits.

Plan	Price	Seats	Max concurrent chips
Free	£0/month	1	3
Pro	£45/month	1	10
Team	£185/month	Unlimited	50

Team self-service is coming soon. C3 can grant Team access to selected accounts before public checkout is enabled.

Manage your plan anytime:

c3 upgrade              # Interactive picker
c3 upgrade pro          # Subscribe to or switch to Pro
c3 upgrade team         # Team is visible but coming soon
c3 upgrade free         # Cancel subscription (takes effect at period end)
c3 upgrade free --yes   # Non-interactive cancel (for scripts)
c3 upgrade free -y      # Short form of --yes

Free to Pro opens Stripe checkout. Downgrades and cancellations take effect at the end of your current billing period. To reverse a pending downgrade or cancellation before it takes effect, run c3 upgrade <current-tier>.

How jobs run

1. PENDING — Job submitted, waiting for compatible hardware
2. SCHEDULING — Assigning to a machine
3. RUNNING — Your script is executing
4. COMPLETED — Done. Download results with c3 pull
5. SYNCED — Results downloaded locally

Failed or cancelled jobs show as FAILED, CANCELED, or TIMED_OUT.

Startup behavior

C3 routes each job to available capacity for the requested hardware profile. Startup time depends on current provider availability, requested hardware, data transfer, and environment setup. GPU classes may hit warm capacity or cold provisioning; CPU profiles currently cold-provision from a C3-managed CPU image.

Providers

By default, c3 deploy asks the C3 marketplace to choose a route before job submission. The control plane then makes the authoritative selection from enabled provider hardware offerings using stock, locality, price, and provider spend guard state.

Use provider: in .c3, c3 deploy -p <provider>, or C3_PROVIDER only when you need to pin a job to a specific provider such as nextgen or nebius. Provider availability is still gated by operator configuration, live inventory cleanup, spend controls, and stock, so a pinned provider can be rejected while auto-routing succeeds on another provider.

Discover public provider IDs and their routable C3 profiles with c3 list --all, or filter with c3 list --provider <provider-id> --all. Public c3 list and /v2/hardware responses expose the provider-to-profile relationship, a coarse availability tier, and the provider-specific C3 billed price. They do not expose regions, native provider SKUs, exact stock, warm-pool capacity, cache affinity, or internal readiness reasons. After submission, c3 deploy prints the matched provider, region, and concrete hardware profile from the created job.

Operators can inspect provider readiness with:

c3 admin balance <provider>
c3 admin provider-status <provider>