NVIDIA
Nemotron 3 Ultra
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...
Cost / 1M tokens
Low usage$0.60
input
$0.12
cached input
$2.40
output
Context Window
262K
Up to 183K max output
Modalities
text → text
Other tokenizer
Quickstart & API
nvidia/nemotron-3-ultra-550b-a55bBest for & Strengths
Key Strengths
Recommended Workloads in ChatOSS
- Coding agents
- Complex reasoning & debugging
- Open-source AI workflows
Verified Evaluations
Benchmarks & Frontier Comparison
How Nemotron 3 Ultra compares against leading open-weight and proprietary frontier models across independent agentic coding and reasoning suites.
| Benchmark (Metric) | Nemotron 3 Ultra | DeepSeek V4.1 Flash | GLM 5.3 | Kimi K3.1 Pro | Claude Opus 5.5 |
|---|---|---|---|---|---|
| Agentic & Software Engineering | |||||
| ChatOSS Agentic Coding Index (%) | — | 63.2% | 58.3% | 65% | 74.1% |
| SWE-Bench Verified (%) | — | 91.5% | 95.4% | 94.8% | 97% |
| DeepSWE v1.1 (Pass@1 %) | — | 77% | 69% | 73.5% | 74.2% |
| Vibe Code Bench v1.1 (%) | — | 87.7% | 78.1% | 88.4% | 90.3% |
| Terminal-Bench 4.0 (%) | — | 42.3% | 42% | 46.2% | 65.2% |
| Code Migration (%) | — | 45.6% | 44.2% | 52% | 66.7% |
| Intelligence & Token Economics | |||||
| Artificial Analysis Intelligence Index (0–100) | 22.9 | 39.5 | 44.8 | 55.1 | 57.6 |
| Input Price (/ 1M tokens) | $0.60 | $0.15 | $1.40 | $2.00 | $4.00 |
| Output Price (/ 1M tokens) | $2.40 | $0.60 | $4.40 | $10.00 | $20.00 |
Artificial Analysis Intelligence Index
Artificial Analysis
Design Arena: 3D
Design Arena · Rank #79 (38.1% win)
Design Arena: ASCIIART
Design Arena · Rank #61 (37.3% win)
Design Arena: Code Generation
Design Arena · Rank #89 (36% win)
Design Arena: Data Visualization
Design Arena · Rank #86 (37.5% win)
Active Provider Endpoints (5)
Upstream inference providers serving Nemotron 3 Ultra, with per-provider pricing, context limits, quantization, and 24-hour uptime.
| Provider | Input / 1M | Output / 1M | Cache Read / 1M | Context | Max Output | Quantization | Uptime (24h) |
|---|---|---|---|---|---|---|---|
DeepInfracache deepinfra/fp4/flex | $0.40 | $1.76 | $0.08 | 262K | 16K | fp4 | 100.0% |
DeepInfracache deepinfra/fp4 | $0.50 | $2.20 | $0.10 | 262K | 16K | fp4 | 99.5% |
BaseTencache baseten/fp4/us | $0.60 | $2.40 | $0.12 | 203K | 183K | fp4 | 78.8% |
BaseTencache baseten/fp4 | $0.60 | $2.40 | $0.12 | 203K | 183K | fp4 | 81.4% |
Venicecache venice/fp8 | $0.625 | $3.13 | $0.188 | 256K | 33K | fp8 | 58.5% |
Supported API Parameters
Parameters accepted by this model when called through ChatOSS or OpenRouter.
Compare Similar Models
Browse all models →NVIDIA
Nemotron 3.5 Super VL
NVIDIA Nemotron 3.5 Super VL is an open multimodal mixture-of-experts model from NVIDIA, with 12B active parameters out of 120B total. It takes text, images, documents, GUI screenshots, and video...
NVIDIA
Switchyard
Switchyard is an open-source model routing system that switches between multiple models to optimize the cost of requests. By default it will use OpenRouter market data to select the most...
NVIDIA
Nemotron 3.5 Lightning
NVIDIA Nemotron 3.5 Lightning is an open mixture-of-experts model from NVIDIA, with 3B active parameters out of 30B total. It is suited for high-throughput agentic workloads and specialized tasks that...
NVIDIA
Nemotron 3.5 Content Safety
NVIDIA Nemotron 3.5 Content Safety is a compact 4B-parameter multimodal guardrail model from NVIDIA, fine-tuned from Google Gemma-3-4B. It moderates both inputs to and responses from LLMs and VLMs, accepting...
Anthropic
Claude Opus 5.5
Claude Opus 5.5 is Anthropic's flagship model for demanding reasoning, coding, and long-horizon agentic work, succeeding Claude Opus 5. It is particularly strong at multi-step changes in large codebases, code...
Moonshot AI
Kimi K3.1 Pro
Kimi K3.1 Pro is Moonshot AI's flagship open-weight multimodal reasoning model and a post-trained successor to Kimi K3, with a 1M-token context window. It is built for long-horizon coding, 3D...