NVIDIA 发布 Nemotron 3.5 Super VL 与 Nemotron 3.5 Bolt 74B-A9B
NVIDIA Releases Nemotron 3.5 Super VL and Nemotron 3.5 Bolt 74B A9B | Artificial Analysis
NVIDIA 发布 Nemotron 3.5 Super VL 与 Nemotron 3.5 Bolt 74B-A9B 两款开源权重模型，前者在 Artificial Analysis 智能指数上得分 26，成为得分最高的美国开源权重模型，超过 Thinking Machines 的 Inkling Small（25.7）、NVIDIA Nemotron 3 Ultra（23）和 Meta Muse Glimmer（17）。Super VL 支持 1M token 上下文与视觉文本输入，参数量 120B/12B，每任务输出约 46k token，比前代 Nemotron 3 Super 少约 60%，在 NVFP4 权重下与 BF16 得分相差 0.1 分以内。Bolt 是新 Bolt 系列首款模型，74B/9B 参数、256k 上下文，智能指数 21.4，在 100B 以下模型中排第 4、美国同级别中排第 1。
正文
当前显示原文，中文全文译文尚未就绪。
Artificial Analysis
All articles
October 15, 2026NVIDIA Releases Nemotron 3.5 Super and Nemotron 3.5 BoltSee model page
Nvidia’s Nemotron 3.5 Super VL is now the leading US open weight model, scoring 26 on the Artificial Analysis Intelligence Index
Nemotron 3.5 Super VL achieves twice the intelligence of its predecessor, Nemotron 3 Super (13), and does so using ~60% fewer output tokens per task
It launches alongside Nemotron 3.5 Bolt 74B-A9B, the first model in its new Bolt family, which scores 21 on the Intelligence Index
Key Takeaways:
➤ Leading US open weight model: Nemotron 3.5 Super VL scores 26.5, edging out Thinking Machines' Inkling Small (25.7) and ahead of NVIDIA's Nemotron 3 Ultra (23) and Meta's Muse Glimmer (17)
➤ Bolt comes within 2 points of Ultra: Nemotron 3.5 Bolt 74B-A9B scores 21, 1.5 points behind Nemotron 3 Ultra 550B A55B (23). It has about a seventh of Ultra's total parameters (74B vs 550B) and a sixth of its active parameters (9B vs 55B)
➤ Super 3.5 near lossless in NVFP4: Nemotron 3.5 Super VL releases in NVFP4, scoring within 0.1 points of its BF16 weights on the Intelligence Index, with no component moving by more than 2.6 points
Nemotron 3.5 Super VL details
Context window: 1M tokens
Modalities: Vision and text input; text output
Parameters: 120B / 12B
Nemotron 3.5 Bolt 74B-A9B details
Context window: 256k tokens
Modalities: Vision and text input; text output
Parameters: 74B / 9B
Nemotron 3.5 Super - US Open Weight Leader
Nemotron 3.5 Super VL is the highest-scoring open weights model from the US, at 26 on the Intelligence Index. It edges out Thinking Machines' Inkling Small (26) and passes NVIDIA's previous flagship, Nemotron 3 Ultra (23), and Meta's Muse Glimmer (17)
Despite the advance, the US still trails the United Arab Emirates (K2 Horizon 375B A23B, 31) and China (MiMo-V2.6-Pro, 46) on open weights intelligence.
Nemotron 3.5 Super - Token Usage
Nemotron 3.5 Super VL uses 46k output tokens per Intelligence Index task, ~60% fewer than its predecessor Nemotron 3 Super (114k) and fewer than K2 Horizon 375B A23B (52k) and MiMo-V2.6-Pro (64k)
Despite a clear increase in token efficiency, Nemotron 3.5 Super still falls a way behind the Intelligence Index vs. Output Tokens per Task frontier
Nemotron 3.5 Super - Output Speed
Measured on a pre-release @DeepInfra endpoint serving NVFP4 weights, Nemotron 3.5 Super VL reaches an output speed of ~630 tokens/s with 100k-token prompts. That is the fastest of any model scoring 20 or more on the Intelligence Index that we have measured at this prompt length
That is ~5x Inkling Small (124 tokens/s), ~2.7x Nemotron 3 Ultra (235), and ~4.5x Nemotron 3 Super (140). It also returns its first token in 1.8s, against 4.4s for Nemotron 3 Ultra and 8.5s for Inkling Small
Nemotron 3.5 Bolt
Nemotron 3.5 Bolt 74B-A9B is the first model in NVIDIA's new Bolt family. It scores 21.4 on the Intelligence Index, placing 4th among models under 100B parameters and first among US models in that class
Bolt scores 8.6 points above Nemotron 3 Super (12.8) with 39% fewer total parameters. Against Nemotron 3 Ultra, it leads on AutomationBench-AA (15.5% vs 3.0%) and AA-Briefcase (912 vs 874 Elo), but trails on AA-LCR (70% vs 79%) and Humanity's Last Exam (25% vs 28%)
Nemotron 3.5 Bolt - Token Usage
Nemotron 3.5 Bolt 74B-A9B uses 82k output tokens per Intelligence Index task, 1.8x Nemotron 3.5 Super VL (46k) and 2.9x Nemotron 3 Ultra (28k), though still 28% fewer than Nemotron 3 Super (114k)
NVIDIA Vera Rubin NVL72 serves up to ~5.7x the total throughput per megawatt of GB300 NVL72 on AA-AgentPerf, with first results on DeepSeek V4 Pro, DeepSeek V4.1 Flash, GLM-5.2 and Kimi K3.
October 15, 2026
First AA-AgentPerf-Local results for an NVIDIA RTX Spark N1X laptop
An NVIDIA RTX Spark N1X laptop completes our local agent workload on Qwen3.8 27B in 18.7 minutes.
October 15, 2026
Welcome to the era of embodied AI: Introducing AA-Robotics v0.1
We are launching AA-Robotics v0.1, our research preview that measures frontier language model performance on understanding, planning, and acting in the physical world. In v0.1, each model is given control of a 6-axis robot arm (5 joints and a gripper) and a simple objective. We’re releasing results for our first eight tasks today.
October 14, 2026