- Context
- 1M
- Reasoning
- Yes
- Included in
- 5 plans
NVIDIA offers a free trial endpoint on build.nvidia.com and does not publish its own per-token API rate. The partner endpoints listed there set their own prices.
API token estimate
Token estimate unavailable
No current Standard API list price is recorded for this exact model. Prices for another model or processing tier cannot supply an estimate.
- Context window
- 1,000,000
- Input
- text
- Output
- text
- Reasoning
- Supported
- Tool calling
- Supported
- Structured output
- Not supported
- Thinking
- Reasoning is configurable on or off with enable_thinking in the chat template.
- API access
- NVIDIA offers a free trial endpoint on build.nvidia.com that records inputs and outputs, which NVIDIA may use to improve its products. NVIDIA does not publish its own per-token rate; listed partner endpoints set their own prices.
- API capabilities
- Function calling is supported. Structured output is listed as not supported.
- Output limit
- A separate maximum output ceiling is not stated; the model card gives a context length of up to 1M tokens.
- Data cutoff
- Pre-training data cutoff September 2025; post-training data cutoff May 2026.
- Open weights
- The weights are downloadable. Use is governed by the OpenMDW License Agreement, version 1.1.
- Official model card checked Sep 29, 2026
- API catalog specifications and access checked Sep 29, 2026
Benchmarks
Open benchmark sheet →No benchmark evidence recorded for this exact release.
Monthly cycle; purchased credits carry over without expiry.
New subscriptions use per-model allowances from September 28, 2026. Existing flat-$10 subscriptions change at their next renewal.
Monthly cycle; purchased credits carry over without expiry.
Monthly cycle; purchased credits carry over without expiry.
Monthly cycle; purchased credits carry over without expiry.
Monthly credit anniversary; included credits do not carry over.
Pricing, assumptions & evidence
- Official model card checked Sep 29, 2026
- Specifications and capabilities checked Sep 29, 2026
- API catalog specifications and access checked Sep 29, 2026
- NVIDIA model card checked Sep 29, 2026
- NVIDIA API catalog model page checked Sep 29, 2026
- Developer-published release day; migrated from reviewed public decision facts checked Sep 29, 2026
Missing token-category prices are not zero. Workload pricing applies exact recorded categories and admitted routes. No benchmark score or quality ranking is inferred from prices.
Aliases, routes and identity
Catalog ID nemotron-3-ultra
Model release (kind: release)
Developer: NVIDIA
nvidia/nemotron-3-ultra-550b-a55bprovider_id
- NVIDIA API catalog model id nvidia/nemotron-3-ultra-550b-a55b checked Sep 29, 2026
nvidia/nemotron-3-ultra-550b-a55bharness_alias