Largest Nemotron 3 model for maximum open-weight reasoning and agent accuracy
https://api.kilo.ai/api/gateway nvidia/nemotron-3-ultra-550b-a55b:free | Benchmark | Score | Metric | Date |
|---|---|---|---|
| SWE-Bench Verified | 70.7 | resolved | 2026-06-04 |
| SWE-Bench Multilingual | 67.7 | resolve rate | 2026-06-04 |
| Terminal-Bench | 56.4 | success rate | 2026-06-04 |
| GPQA | 87 | accuracy | 2026-06-04 |
nemotron-3-ultra-550b-a55b is listed for chat, reasoning workloads and supports a 1.0M context window.
Use it when Kilo Code's free tier is enough for evaluation, demos, or light production traffic.
We found 3 provider listings for nemotron-3-ultra-550b-a55b. Check model ID, quota, pricing, and API format before switching providers.
| Provider | Model listing | Access | Context | API | Limits |
|---|---|---|---|---|---|
| | nvidia/nemotron-3-ultra-550b-a55b:free | Free tier | 1.0M | OpenAI-style | ~200 req/hr |
| | NVIDIA: Nemotron 3 Ultra (free) | Free tier | 1.0M | OpenAI-style | 200 req/day (free tier) |
| | Nemotron 3 Ultra 550B A55B | Free tier | 1.0M | Native | Varies |
| Model | Provider | Context | Access |
|---|---|---|---|
| NVIDIA: Nemotron 3 Ultra (free) | OpenRouter | 1.0M | Free tier |
| Nemotron 3 Ultra 550B A55B | NVIDIA NIM | 1.0M | Free tier |
| stepfun/step-3.7-flash:free | Kilo Code | 262K | Free tier |
| nvidia/nemotron-3-super-120b-a12b:free | Kilo Code | 262K | Free tier |
| nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free | Kilo Code | 256K | Free tier |
nemotron-3-ultra-550b-a55b is tagged for chat in this catalog and works with OpenAI-compatible client libraries.
nemotron-3-ultra-550b-a55b is tagged for reasoning in this catalog and works with OpenAI-compatible client libraries.
nemotron-3-ultra-550b-a55b is listed with free API access on Kilo Code, subject to the provider's quota and account policy.
The model ID shown in this catalog is nvidia/nemotron-3-ultra-550b-a55b:free.
The listed free tier limit is ~200 req/hr. Limits can change per account tier, so confirm against the provider dashboard.
The listed context window is 1.0M tokens with up to 65K output tokens.
NVIDIA Nemotron 3 Ultra is an open frontier-reasoning and orchestration model from NVIDIA, with 55B active parameters out of 550B total (MoE). Built on a hybrid Transformer-Mamba mixture-of-experts architecture, it... **Terms of service** For NVIDIA free endpoints (Super/Ultra/etc): Trial use only - do not submit personal or confidential data. Your use is logged for security purposes and to improve NVIDIA products and services. The logged session data for improvement purposes is not linked to your identity or any persistent identifier. For more information about our data processing practices, see our [Privacy Policy](https://www.nvidia.com/en-us/about-nvidia/privacy-policy/). By interacting with this endpoint, you consent to our collection, recording, and use of such information and the [NVIDIA API Trial Terms of Service](https://assets.ngc.nvidia.com/products/api-catalog/legal/NVIDIA%20API%20Trial%20Terms%20of%20Service.pdf).
For API keys, setup steps, and provider-level limits, see the Kilo Code provider page.