Ask
Launch

Modal adds NVIDIA B300 GPU and separately-priced Sandbox + Notebooks compute

Modal pricing

Modal added the B300 (Blackwell Ultra) GPU at $0.001972/sec above the B200, plus a separately-metered Sandbox + Notebooks tier at ~3x standard CPU/memory rates, and now itemizes region (1.5-1.75x) and non-preemptible (3x) modifiers.

Before

Top GPU SKU was the B200 at $0.001736/sec; no dedicated Sandbox/Notebooks rate card; region and non-preemptible multipliers not itemized on the pricing page.

After

B300 available at $0.001972/sec; Sandbox + Notebooks compute metered at $0.00003942/core/sec CPU and $0.00000667/GiB/sec memory; region selection billed 1.5-1.75x base and non-preemptible execution 3x base across all plans.

Modal extended its per-second rate card at the high end with the NVIDIA B300 (Blackwell Ultra) at $0.001972/sec, sitting above the B200 ($0.001736/sec) as the platform new most-expensive GPU SKU. All existing GPU, CPU, memory, and storage rates and the plan fees (Starter $0, Team $250 + compute, Enterprise custom) were unchanged.

Two structural additions accompany the new SKU. First, a dedicated Sandbox + Notebooks compute rate card now meters interactive/burst workloads separately at $0.00003942/core/sec CPU and $0.00000667/GiB/sec memory — roughly triple the standard function rates — while GPU on sandboxes still follows the standard per-second card. Second, the pricing page now explicitly itemizes two per-second modifiers that apply across Starter, Team, and Enterprise: region selection at 1.5-1.75x base prices and non-preemptible (guaranteed, non-interruptible) execution at 3x base prices.

From Modal's pricing timeline
B300 GPU + Sandbox/Notebooks Pricing + Rate Modifiers

Modal added the NVIDIA B300 (Blackwell Ultra) at $0.001972/sec — its new top-end GPU SKU above the B200 ($0.001736/sec). A separately-metered Sandbox + Notebooks compute tier appeared at $0.00003942/core/sec CPU and $0.00000667/GiB/sec memory (roughly 3× the standard function rates, reflecting burst-oriented interactive workloads). The pricing page also now itemizes two per-second modifiers: region selection at 1.5–1.75× base prices and non-preemptible (guaranteed) execution at 3× base prices. All existing GPU, CPU, memory, storage rates and plan fees ($0 / $250 / Custom) were unchanged.

About Modal
modal.com ↗

Modal runs a per-second pure-usage compute model: GPU rates from T4 at $0.000164/sec to B300 at $0.001972/sec, with B200 at $0.001736/sec, H200 at $0.001261/sec, H100 at $0.001097/sec, RTX PRO 6000 at $0.000842/sec, A100 80GB at $0.000694/sec, A100 40GB at $0.000583/sec, L40S at $0.000542/sec, A10 at $0.000306/sec, and L4 at $0.000222/sec.

Free tier
Yes
Commits
Available
Transparency
public

Modal pricing history

  1. Aug 2026
    Shared API Access Narrowed to Team and Enterprise
  2. Aug 2026
    Team Plan Container Concurrency Raised 5x to 5,000
  3. Jul 2026
    Shared API Launched — Token-Based Pricing Alongside Per-Second GPU
  4. Jul 2026
    B300 GPU + Sandbox/Notebooks Pricing + Rate Modifiers
  5. Feb 2026
    AWS + GCP Marketplace Billing for Enterprise
Full Modal timeline

More Modal activity

All pricing activity