Nemotron 3 Ultra
NVIDIAReleased Jun 4, 2026Open weights · NVIDIA Open Model Licensenemotron-3-ultra-550b-a55b
- Intelligence (Epoch)
- 146 · 58th of 111
- Reasoning
- yes, with effort levels
- Knowledge cutoff
- June 2024
- Context
- 262K tokens
- Max output
- 32k tokens
- Input
- text
- API features
- tools, structured outputs, JSON mode, seed
Read Oct 23, 06:00 UTC from the hosts' published offers and OpenRouter, spec sheet from NVIDIA · cutoff: aiknowledgecutoff.com
Free right now
- Free OpenRouter, served by Nvidia · no end date · 20 req/min, 50 req/day, 1,000 req/day with $10 of credits
Checked Oct 23 · every free model on Deals
The cheapest host, DeepInfra, serves fp4 weights (maker: bf16). Weights and context per host
Intelligence for the price
17 models score higher for the same price or less; it ranks 18 of the 64 models at its price or below.
Epoch Capabilities Index against the lowest input + output price per M, log scale · prices Oct 23. Score: Epoch AI, CC BY 4.0, 146 with an interval of 144 to 148.
Against its rivals
The best of its price range and the top score
- Nemotron 3 Ultrathis model$2.70146
- Claude Opus 5.5top score$24.00169
- Gemini 3.7 Flashsame price range$4.50158
- Gemini 3.8 Flashsame price range$4.50157
Same level for 17× less: Qwen3.7 Flash at Alibaba, $0.16 per M input + output · their Epoch intervals overlap (144 against 146).
Bars: Epoch Capabilities Index (Epoch AI, CC BY 4.0), drawn from 100, below the lowest score in the catalogue · price: lowest input + output per M.
Speed by provider
Output speed of Nemotron 3 Ultra at each host, fastest first
- DeepInfra82 tok/s1.55 s to first token
- BaseTen82 tok/s0.32 s to first token
- Venice56 tok/s1.47 s to first token
Speed: OpenRouter, median of the last 30 min, read Oct 23, 06:00 UTC · a snapshot of real traffic before the daily read, not a daily average · bars from zero.
Run it yourself
open weights, NVIDIA Open Model License · too large for the 6 workstations in our guideSelf-host guide ›- Parameters
- 550B
- 55B active per token (MoE)
- Memory needed
- 330 GB
- in INT4 · 660 GB in INT8
- Smallest cloud setup
- 8× H100
- 640 GB · $17.20/h, about $12,556 a month
- Worth it above
- 5B tokens
- a month, against the cheapest API at $2.70 per M
| Quantization | Memory | Smallest cloud GPU | Quality |
|---|---|---|---|
| INT4 (Q4) | 330 GB | 8× H100 · $17.20/h | smallest footprint, minor quality loss |
| INT8 (Q8) | 660 GB | multi-node, over 640 GB | near-lossless |
| FP16 (full) | 1320 GB | multi-node, over 640 GB | reference quality |
At home: 330 GB in INT4 is beyond any desktop or Mac in our machines table, so it means a multi-GPU server.
Memory ≈ parameters × bytes per parameter × 1.2 · cloud price: on-demand rate, running 24/7 · builds: GGUF, AWQ, MLX, GPTQ, Ollama · break-even calculator
| Provider | Input | Output | Cache ↑ | tok/s | First token | Uptime 24h | Checked |
|---|---|---|---|---|---|---|---|
| DeepInfrafp4 (bf16) · 256K ctx · 16K out | $0.50 | $2.20 | $0.10 | 82 | 1.55 s | 100% | Oct 23 |
| BaseTenfp4 (bf16) · 203K ctx (256K) | $0.60 | $2.40 | $0.12 | 82 | 0.32 s | 99.9% | Oct 23 |
| Venicefp8 (bf16) · 250K ctx · 32K out | $0.63 | $3.13 | $0.19 | 56 | 1.47 s | 98.7% | Oct 23 |
3 of 3 providers · prices per million tokens, read Oct 23, 06:00 UTC · speed: OpenRouter, median of the last 30 min, read Oct 23, 06:00 UTC · uptime: OpenRouter, last 24 h · under each host: weights precision, context and max output as served, the maker's value in brackets when the host's differs
| Provider | Weights | Context | Max output | Tools | Structured outputs | Reasoning |
|---|---|---|---|---|---|---|
| DeepInfra | fp4 | 256K | 16K | yes | yes | yes |
| BaseTen | fp4 | 203K | not checked yet | yes | no | yes |
| Venice | fp8 | 250K | 32K | yes | no | yes |
Weights, context and max output: read at each host's own catalogue or docs where its terms allow, otherwise as the host declares to OpenRouter; hover a cell for its source and date · tools, structured outputs, reasoning: as declared to OpenRouter · not verified by us
Nemotron 3 Ultra spec sheet, field by field, with sources
Nemotron 3 Ultra spec sheetmaker-declared
What NVIDIA declares about the model itself — read from its own weights repository and documentation, not from any host. What each endpoint actually serves is in the pricing above: the two can differ, and both are true.
- Context window
- 262,144 tokens
- Parameters
- 560,524,578,816≈ 560.5B
- Active experts
- 22 of 512 active per token
- Weight precision
- BF16
- License
- NVIDIA Open Model License — open weights, conditionalcuratedScope: the licence on the published weights, and nothing else. It is not the maker's acceptable-use policy, which is a separate document, and it is not the contract you sign with whichever host you call — hosts set their own terms, which this index does not read.
- Knowledge cutoff
- not published by the maker≈ June 2024 third-party estimate — not the maker's figurecurated
- Modalities
- text
- Max output (official)
- to verify
- Release date
- to verify≈ June 4, 2026 third-party catalogue — not the maker's figure
- Maker lifecycle status
- to verify
Also declared in the weights
- Architecture class
- NemotronHForCausalLM
- Vocabulary
- 131,072 tokens
- Weights first published
- June 3, 2026
- Weights last updated
- August 24, 2026
- Languages declared
- 12 languages
nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 · config.json @ 77df655 · read 2026-10-08 · 6 of 10 fields published by the maker
| Date | Provider | Input | Output | Change |
|---|---|---|---|---|
| Oct 22 | DeepInfra | $0.40 → $0.50 | $1.76 → $2.20 | +25% |
| Oct 16 | DeepInfra | $0.50 → $0.40 | $2.20 → $1.76 | −20% |
Every published change of input or output price, newest first · change on input + output · full series: pricing-history.csv.
| Provider | Chatbot 1,000 in / 500 out | RAG / long context 8,000 in / 500 out | Batch summarize 4,000 in / 1,000 out |
|---|---|---|---|
| DeepInfra | $1.60 | $5.10 | $4.20 |
| BaseTen | $1.80 | $6.00 | $4.80 |
| Venice | $2.19 | $6.56 | $5.62 |
What each host bills for 1,000 requests of a representative job
Your own workload
Also serving Nemotron 3 Ultra, without a published price
No price we can verify, so no rank.
Other NVIDIA models · all its models, hosts and history ›
- Nemotron 3.5 Super VL7 providers · from $0.50
- Nemotron 3 Nano 30B A3B4 providers · from $0.25
- Nemotron 3 Super4 providers · from $0.48
- Nemotron 3.5 Lightning4 providers · from $0.18
- Nemotron 3.5 Content Safety1 provider · from $0.40
- Nemotron 3 Nano Omni 30B A3B Reasoning0 provider · from —