🔍 Read the full analysis: Nemotron 3.5 Super VL Takes Less Than A Second Per Task on ThorstenMeyerAI.com
Get business pricing on monitors, keyboards and dev gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
NVIDIA released Nemotron 3.5 Super VL on October 15, 2026. Artificial Analysis rated it 26 on its Intelligence Index and measured high speed on one endpoint, while provider-wide results vary; the source says it costs $0.25 per index task and trails several similarly priced models on the score.
The model has 120 billion total parameters, with 12 billion active in its mixture-of-experts design, and a listed one-million-token context window. Its weights are released in NVFP4, a four-bit format. The source says its OpenMDW-1.1 license allows commercial use. Artificial Analysis lists API prices of $0.20 per million input tokens and $0.575 per million output tokens, with cached input at $0.055 per million tokens.
Artificial Analysis puts the cost of one Intelligence Index task at $0.25. The source compares that with $1.64 for Nemotron 3 Super, the predecessor, which scored 13. The new model’s lower task cost is attributed in the source to fewer output tokens per evaluation and discounted cached input. In the reported cost breakdown, about 89% of the task bill is input, including cache writes; output costs less than three cents.
Speed depends on the serving provider. On a pre-release DeepInfra endpoint using NVFP4 weights, the model generated about 630 output tokens per second with prompts of 100,000 tokens. Artificial Analysis’ figures across six providers show a median of 196 tokens per second on its standard prompt and 307 with 100,000-token prompts, with a range from 106 to 534 at the reported percentiles. Median time to first token was 0.89 seconds; the source reports about 11 seconds before the first answer token when reasoning is included, and 13.6 seconds for a 500-token answer.
Infographic 01 · What 25 cents buys
Cheap and fast. Not the most intelligence per dollar.
Nemotron 3.5 Super VL costs $0.25 per Intelligence Index task and scores 26. Several open models with MIT licenses score 13 to 20 points more for the same money or less.
Intelligence Index vs. cost to run it
Up and to the left is better. The dashed line marks Super VL’s $0.25. Hover a dot for license and values.
Not plotted: K2 Horizon 375B A23B (UAE, index 30.5, listed at $0.00) and Nemotron 3.5 Bolt 74B-A9B (index 21.4, no API price).
License matters here. Muse Spark Open (Meta, 38.9) scores higher than Super VL but ships under the Muse Model License, which allows commercial use with restrictions. Super VL’s OpenMDW-1.1 license allows commercial use.
Show the data as a table (20 models)
| Model | Creator | License | Index | Cost / task |
|---|---|---|---|---|
| MiMo-V2.6-Pro | Xiaomi | MIT | 46.3 | $0.13 |
| GLM-5.3 | Z AI | GLM-5.3 License (restricted) | 44.8 | $2.01 |
| Step 5 | StepFun | StepFun Model License (restricted) | 43.7 | $1.03 |
| Kimi K3 | Kimi | Kimi K3 License (restricted) | 43.6 | $2.00 |
| Claude Haiku 5.5 (max) | Anthropic | Proprietary | 43.4 | $0.21 |
| GLM-5.3-Flash | Z AI | MIT | 41.8 | $0.28 |
| Ling 3.1 Flash | InclusionAI | MIT | 41.1 | $1.00 |
| Qwen3.8 Flash Next | Alibaba | Qwen Community License (restricted) | 39.8 | $0.37 |
| DeepSeek V4.1 Flash | DeepSeek | MIT | 39.5 | $0.27 |
| Muse Spark Open | Meta | Muse Model License (restricted) | 38.9 | $0.84 |
| GPT-6 Luna (max) | OpenAI | Proprietary | 38.1 | $0.07 |
| MiMo-V2.6-Flash | Xiaomi | MIT | 37.9 | $0.06 |
| DeepSeek V4 Pro | DeepSeek | MIT | 36.0 | $0.67 |
| Qwen3.8 27B | Alibaba | Apache 2.0 | 33.7 | $1.01 |
| MiniMax-M3 | MiniMax | MiniMax Community License (restricted) | 29.2 | $0.51 |
| Nemotron 3.5 Super VL | NVIDIA | OpenMDW-1.1 | 26.5 | $0.25 |
| Inkling Small | Thinking Machines | Apache 2.0 | 25.7 | $0.09 |
| Nemotron 3 Ultra 550B | NVIDIA | OpenMDW | 22.9 | $0.60 |
| Muse Glimmer | Meta | Apache 2.0 | 17.5 | $0.05 |
| Nemotron 3 Super (previous) | NVIDIA | NVIDIA Open Model License | 12.8 | $1.64 |
Source: Artificial Analysis, Nemotron 3.5 Super VL model page and model data for the comparison models, read 16 Oct 2026; launch analysis (15 Oct 2026). Intelligence Index v4.3.2, ten evaluations. Cost per task = total API cost to run the Intelligence Index divided by its tasks. “Restricted” = Artificial Analysis lists commercial use as allowed with conditions.
Powered by Thorsten Meyer AI
Speed Comes With a Score Trade-Off
For teams running high-volume document or repository workflows, the low reported task cost and discounted cached input may matter as much as raw benchmark performance. A one-million-token context window also gives operators room to supply large inputs, though actual costs will depend on how much they send and how effectively caching applies.
The source’s comparisons show why the score alone is not a value verdict. It lists MiMo-V2.6-Pro at 46 for $0.13 per task, DeepSeek V4.1 Flash at 39 for $0.27, and Inkling Small at 26 for $0.09. These figures suggest Super VL’s case rests on other considerations, including endpoint speed, commercial licensing and the option to run weights on NVIDIA hardware. The source does not provide a common deployment-cost comparison for self-hosting.
Latency could limit use in interactive applications. Although the median first-token measurement is under a second, the source says the model reasons before producing its answer, with the first answer token arriving after about 11 seconds. That may suit background agents and pipelines better than a chat product where users expect a quick substantive response.
As an affiliate, we earn on qualifying purchases.
How the Benchmark Rankings Work
Artificial Analysis assigns Nemotron 3.5 Super VL an Intelligence Index score of 26.5 in the source’s specification table, rounded to 26 elsewhere in the article. Its first-place position applies to 67 open-weight models in the 40-billion-to-150-billion total-parameter class. Across all open-weight models, the source places it around 18th. The size-class ranking therefore describes a narrower comparison group than the full open-weight field.
The source also reports gains over Nemotron 3 Super on selected evaluations. GDPval-AA, which measures economically useful work, rose to 1,162 Elo from 490. AA-Briefcase reached 999 Elo, compared with 876 for Nemotron 3 Ultra. AutomationBench-AA registered 14.8%, versus 3.0% for Ultra; the source notes that the newer model’s figure remains low in absolute terms.
On AA-Omniscience, the source reports a 36% hallucination rate, down from 87% for Nemotron 3 Super. It also compares the model with NVIDIA’s predecessor on task cost and score. These are benchmark results cited by the source, not guarantees of performance on a particular organization’s data or applications.
large language model developer tools
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Provider Results Still Vary
The speed figures are not uniform across serving providers: the 630-token-per-second result came from one pre-release endpoint, while the six-provider median reported by Artificial Analysis was lower. The source advises measuring the endpoint a team plans to use. It does not identify how the provider range would translate into performance under every workload or service configuration.
The supplied material does not detail the test methodology, uncertainty intervals, or how the benchmark results map to specific production tasks. It also does not provide a full self-hosting cost estimate, including hardware, power and operations. Those gaps make it difficult to compare the total cost of self-hosting with API use based on the listed per-task figure alone.
The source says the model is the top-scoring permissively licensed U.S. open-weight model, but the comparison depends on how licenses are classified. It contrasts Super VL with Meta’s Muse Spark Open, whose license it describes as allowing commercial use with restrictions. The practical implications of those terms depend on the intended use and legal review.
As an affiliate, we earn on qualifying purchases.
Measure Before Deployment
Operators evaluating the model can compare it with alternatives on their own prompts, latency targets and input patterns, including workloads with repeated documents where cached input may affect costs. They should also verify current provider pricing, endpoint speed and license terms before committing to a deployment.
Further provider measurements and real-world evaluations will help clarify whether the reported speed and task-cost advantages hold across production settings. The supplied source material gives no schedule for additional benchmark updates or a separate next release milestone.
As an affiliate, we earn on qualifying purchases.
Key Questions
When was Nemotron 3.5 Super VL released?
NVIDIA released Nemotron 3.5 Super VL on October 15, 2026, according to the source. Artificial Analysis published benchmark figures the same day.
What Intelligence Index score did it receive?
The source reports a score of 26.5, rounded to 26. Its first-place ranking applies to the specified open-weight size class of 40 billion to 150 billion total parameters.
How much does an Intelligence Index task cost?
Artificial Analysis lists a cost of $0.25 per index task. The source says about 89% of that example cost comes from input, including cache writes; actual bills depend on workload and provider.
How fast is the model?
A pre-release DeepInfra endpoint produced about 630 output tokens per second with 100,000-token prompts. Across six providers, the reported median was 307 tokens per second at that prompt length, so results vary by endpoint.
Does the license allow commercial use?
The source identifies the license as OpenMDW-1.1 and says commercial use is allowed. Teams should check the license terms against their planned use.
Source: ThorstenMeyerAI.com
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
