🔍 Read the full analysis: Nemotron 3.5 Super VL: Fast Latency And A Score Of 26 on ThorstenMeyerAI.com
Get smart everyday buys delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
NVIDIA released Nemotron 3.5 Super VL on October 15, 2026. Artificial Analysis gave it an Intelligence Index score of 26.5; the source report highlights fast inference, lower task costs than its predecessor and a commercial-use license, while noting that several competing models score higher for similar or lower costs.
NVIDIA released Nemotron 3.5 Super VL on October 15, and Artificial Analysis rated it at 26.5 on its Intelligence Index. The model combines an open-weight release and a license allowing commercial use with reported high inference speeds, but the score is well below several models available for a similar or lower estimated cost per benchmark task.
The model has 120 billion total parameters, with 12 billion active in its mixture-of-experts design. NVIDIA released weights in NVFP4, a 4-bit format, and lists a one-million-token context window with text and image input and text output. The source report identifies the license as OpenMDW-1.1 and says it permits 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.
Artificial Analysis estimates $0.25 per Intelligence Index task, compared with $1.64 for the previous Nemotron 3 Super. The report attributes the lower figure in part to fewer output tokens per benchmark task and inexpensive cached input; its estimate says roughly 89% of the new model’s task cost comes from input, including cache writes. These are benchmark-based estimates, not a guarantee of what a particular deployment will cost.
Speed results vary by measurement. A pre-release DeepInfra endpoint running the NVFP4 weights produced about 630 output tokens per second on 100,000-token prompts, according to the report’s account of Artificial Analysis measurements. Across six providers, the model page showed a median of 307 tokens per second at that prompt length, with results ranging from 106 to 534. Median time to first token was 0.89 seconds, while the report says reasoning can delay the first answer token to about 11 seconds.
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 Value Trade-Off
Nemotron 3.5 Super VL’s case is not simply that it offers the most intelligence for the price. The source compares its 26.5 score and $0.25 estimated task cost with models that score higher for similar or lower benchmark costs. For example, it reports MiMo-V2.6-Pro at 46 for $0.13 per task and Inkling Small at 26 for $0.09. Those figures come from the cited benchmark comparison and may not match costs for every workload or provider.
For teams, the more relevant advantages may be inference speed, commercial-use licensing and the option to run the weights on their own NVIDIA hardware. The reported gap between one endpoint’s speed and the multi-provider median also means performance depends on where the model is served. Buyers should test their intended endpoint and workload rather than treating a single speed result as universal.
The index score also conceals uneven results across tasks. The source says GDPval-AA, a measure of economically useful work, rose to 1,162 Elo from 490 for Nemotron 3 Super, and AutomationBench-AA reached 14.8%. It also reports a drop in hallucination rate on AA-Omniscience from 87% to 36%. Those are specific benchmark results, not evidence that the model will perform equally well in every business process.
Popular right now
KitchenAid Commercial 8-Quart Bowl-Lift Stand Mixer, White
Samsung 50-inch QN70H Neo QLED 4K Smart TV with Austere 2.5 m 8K HDMI Cable
Bose QuietComfort Wireless Noise-Cancelling Headphones, Twilight Blue
As an affiliate, we earn on qualifying purchases.
How the Score Is Ranked
Artificial Analysis places Nemotron 3.5 Super VL first among 67 open-weight models in its 40-billion-to-150-billion-parameter size class, according to the source report. That is a narrower comparison than a ranking across all open-weight models: the report estimates the model is about 18th when models of every size are counted once. The distinction matters when interpreting the “number one” claim.
The report says the new model’s Intelligence Index score is about twice the 13-point score of Nemotron 3 Super. It also says NVIDIA’s model leads among permissively licensed US open-weight models. That claim depends on how the model set and licenses are defined. The source contrasts it with Muse Spark Open, which scored 39 but uses a license described as allowing commercial use with restrictions.
““Super VL is a very good engine and an average brain.””
— Thorsten Meyer, ThorstenMeyerAI.com
Provider Results and Real-World Performance
The figures do not establish how Nemotron 3.5 Super VL will perform across all deployment settings. The 630-token-per-second result came from one pre-release endpoint, while the reported median across six providers was lower. The source does not give enough detail here to establish that every provider used identical hardware, configurations or serving conditions.
Benchmark task costs and scores are also estimates tied to Artificial Analysis methods. Actual spending will depend on prompt length, caching, output volume, provider rates and workload. The source describes gains in knowledge work, SaaS automation and hallucination testing, but does not establish performance across all tasks or independently verify the model’s behavior in production.
Test the Intended Deployment
The next practical step for teams considering the model is to test it on their own prompts, documents and software workflows, using the provider and hardware they expect to deploy. Those tests can reveal whether the model’s response latency, output quality and input-heavy costs fit a particular application.
Further provider measurements and independent evaluations would help clarify how consistent the speed and benchmark results are. Until then, the confirmed release and published scores make Nemotron 3.5 Super VL a model to compare—not a universal recommendation based on one headline score or one endpoint result.
Key Questions
What score did Nemotron 3.5 Super VL receive?
Artificial Analysis scored it at 26.5 on its Intelligence Index. The source report rounds this to 26 in its headline and says the model ranked first among open-weight models in its 40-billion-to-150-billion-parameter size class.
How fast is the model?
A pre-release DeepInfra endpoint reportedly produced about 630 output tokens per second on 100,000-token prompts. Across six providers, the reported median at that prompt length was 307 tokens per second, so speed depends on the endpoint.
How much does a benchmark task cost?
Artificial Analysis estimated $0.25 per Intelligence Index task. This is a benchmark estimate, not a fixed charge for every use; actual costs depend on provider pricing and the volume of input, cached input and generated output.
Can businesses use the model commercially?
The source identifies the model’s license as OpenMDW-1.1 and says it allows commercial use. Organizations should review the license terms themselves before deployment.
Does a score of 26 make it the best-value open model?
Not according to the comparisons in the source report. It lists several open-weight models with higher Intelligence Index scores at similar or lower estimated task costs. Nemotron’s case instead includes speed, licensing and the option to run the weights on NVIDIA hardware.
Source: ThorstenMeyerAI.com
