Provider Dossier · GPU Cloud
Modal
Code-first serverless GPUs; full control of the serving stack.
The Datum Index verdict
Highly commended
Ranked No. 7 of 20 in our composite reputation index, and currently improving.
≈ 4.3 / 5 aggregate · 300 reviewer notes
Composite index · reviewed 10 July 2026
Findings
- Modal holds a Datum Index composite reputation of 85 / 100 (≈ 4.3 / 5), ranked No. 7 of 20.
- It is a gpu cloud based in New York, USA, founded 2021, offering open-weight model access.
- Its flagship offering is Serverless GPU runtime, through its own API.
- Pricing is usage-metered rather than per-token (medium confidence).
- Reference throughput is ~156 tok/s with ~2000 ms time-to-first-token; stated uptime is 99.7%.
Modal enters our index at No. 7, a placement that reads as highly commended. The composite rests on 300 reviewer notes gathered across the public record, and it is best understood not as a score out of ten but as a standing among peers. The trend line is pointed upward: reviewer sentiment has strengthened over recent quarters.
In its favour
Code-first serverless GPUs; full control of the serving stack.
On the numbers that flatter it: open-weight access, 2 named compliance attestations, and a reference throughput of ~156 tokens per second.
Points of concern
You build the inference layer; not a turnkey token API.
The caveat a buyer should price in before committing production traffic. As with every entry, we weigh it against the field rather than against perfection.
Incident & Uptime Note
The record of being there
Modal states an availability of 99.7%. Held over a full year, that headline implies on the order of 26.3 hours of cumulative unavailability — a useful sense of scale, though real incidents cluster rather than spread evenly, and a single bad afternoon can outweigh a quiet quarter.
Observatory Panel
Latency & throughput, in distribution
Monthly panels across the models Modal serves medium
Kimi K3 (self-served)
| Window | TTFT ms — mean · p50 · p90 | Throughput tok/s — mean · p50 · p90 | ||||
|---|---|---|---|---|---|---|
| May 2026 | ~2390 | ~2120 | ~4250 | ~140 | ~147 | ~172 |
| June 2026 | ~2080 | ~1950 | ~3790 | ~151 | ~160 | ~186 |
| July 2026 | ~2330 | ~2060 | ~4340 | ~148 | ~151 | ~178 |
Llama 3.3 70B (self-served)
| Window | TTFT ms — mean · p50 · p90 | Throughput tok/s — mean · p50 · p90 | ||||
|---|---|---|---|---|---|---|
| May 2026 | ~1670 | ~1570 | ~2920 | ~60 | ~62 | ~70 |
| June 2026 | ~1720 | ~1620 | ~2760 | ~56 | ~60 | ~70 |
| July 2026 | ~1520 | ~1410 | ~3010 | ~65 | ~69 | ~77 |
All figures are single-stream, per-request rates as one user would see them — never aggregate accelerator throughput. Serving modes are not comparable with one another: shared serverless APIs batch many tenants per accelerator, dedicated and self-served deployments hand one tenant the whole card, and specialist silicon is a regime of its own. Read each figure within its mode.
No field returns on record for Modal yet — the ledger above opens as soon as the first practitioner submission clears verification.
Key Facts
The record, in brief
| Category | GPU Cloud |
|---|---|
| Headquarters | New York, USA |
| Founded | 2021 |
| Model access | Open-weight |
| Flagship / reference | Serverless GPU runtime · Bring-your-own; billed by GPU-second |
| OpenAI-compatible API | No |
| Pricing model | Usage-metered (per GPU-second) medium |
| Reference throughput | ~156 tok/s |
| Reference TTFT | ~2000 ms |
| Stated uptime | 99.7% |
| Compliance | SOC 2 Type II · GDPR |
| Composite reputation | 85 / 100 · ≈ 4.3 / 5 · 300 notes |
Reviewer Notes
What the field says
The 300 notes behind Modal's standing are drawn from the accumulated public verdict of practitioners — the recurring praises and the recurring gripes — rather than from any single survey. Two themes dominate. Admirers return to one point above all: Code-first serverless GPUs; full control of the serving stack. Detractors return to another: You build the inference layer; not a turnkey token API.
We publish neither individual reviews nor reviewer identities; the composite is an editorial synthesis, and the reviewer-note count is an order-of-magnitude indication of how much public signal informs it.
Related dossiers · see also
Peer entries in the same or adjacent category, for comparison against Modal's standing.