Nemotron 3.5 Lightning Tested: 670 tokens/s Speedster for Agents
ArtificialAnlys · x · 2026-08-11
According to tests by Artificial Analysis, NVIDIA's newly released Nemotron 3.5 Lightning is an open-weights small model focused on extreme efficiency. With 31.6B total parameters (3.6B active), it achieves a dual leap in intelligence and speed while maintaining a compact footprint.
Key Data & Highlights:
- Major Intelligence Jump: Scores 24 on the Artificial Analysis Intelligence Index, a +9 point improvement over its predecessor (15), matching the larger gpt-oss-120b.
- Extreme Inference Speed: Achieves a blazing 670 tokens/second output speed on DeepInfra's NVFP4 quantized endpoint, vastly outperforming peers in its class.
- Agentic Step Change: Scores 24% on Terminal-Bench v2.1 (up from 7% previously) and surpasses the larger Nemotron 3 Super on GDPval-AA v2.
- Open Source & Availability: Released under the permissive OpenMDW-1.1 license for commercial use, features a 1M token context window, and is available across multiple inference providers.
Related event: NVIDIA Unveils Nemotron 3.5 Lightning and NeMo Switchyard Router(55 posts)→
More from Models
- Meta open-sources Muse Spark Open agent model with day-0 TPU support in SGLang — BanghuaZ · 2026-10-16
- Saying "load-bearing" in your prompt makes Opus 5.5 stop holding back — zacharynado · 2026-10-16
- Americans get their Claude accounts banned while on business trips in China — dotey · 2026-10-16
- I priced every Grok Bot call: $20 to $200 plan in 5 days, long jobs eat 52% — jonathanmalkin · 2026-10-16
- NVIDIA releases Nemotron-3.5-Embed-1B, a multimodal embedding model covering 34 languages — jacek2023 · 2026-10-16
- EpochAI pegs GPT-6 training at up to 1.8e27 FLOP; scaling01 calls param estimates way too high — scaling01 · 2026-10-16