NVIDIA AI & NIMopen-weight
Released: 2026-08-11Artificial AnalysisNemotron 3.5 Lightning
NVIDIA Nemotron 3.5 Lightning is an open mixture-of-experts model from NVIDIA, with 3B active parameters out of 30B total. It is suited for high-throughput agentic workloads and specialized tasks that...
Intelligence Index
5.5/ 100
Artificial Analysis Intelligence Index (independent composite)Throughput Speed
269tok/sec
Output streaming throughputTime to First Answer
13s
Time to first answer token, including any thinkingCost to Run the Index
0.31USD
Artificial Analysis cost of the full Index runMeasured results · Artificial Analysis
Independent evaluations of Nemotron 3.5 Lightning. Values are reported as published; nothing here is estimated or derived. Snapshot 2026-11-15.
- Intelligence Index
- 5.5
- Cost to run the Index
- $0.31
- Time per task
- 4.0 min
- Omniscience index
- -15.7
- Reasoning model
- Yes
- Cached-input price
- $0.04 / 1M
Reasoning & knowledge
- AA-Omniscience Accuracy14.7%
- CritPt0.0%
- CritPt v2 (Hard)1.5%
- GPQA Diamond74.3%
- HLE-Diamond4.5%
- Humanity's Last Exam10.6%
Coding & terminal agents
- FrontierCode-AA 2.00.0%
- SciCode32.1%
- Terminal-Bench 2.124.3%
- Terminal-Bench 4.00.5%
- Terminal-Bench-Science 0.10.0%
Agentic work
- AA-Briefcase v1.1 (Elo)507 Elo
- AutomationBench-AA0.8%
- EnterpriseOps-Gym-AA18.4%
- GDPval-AA (normalized Elo)645 Elo
- τ³-Banking8.9%
Long context & documents
Health & safety
Headline Evaluation Breakdown
Six independently measured evaluations (Artificial Analysis)Reasoning · CritPt v2 (Hard)1.5%
Coding & agents · Terminal-Bench 4.00.5%
Long context · AA-LCR60.3%
Knowledge · Humanity's Last Exam10.6%
Sourced Benchmark Evaluations (18)
Verified performance across authoritative benchmarks with provenance tracking.| Benchmark | Status | Score | Trust | Date | Source Type | Provenance |
|---|---|---|---|---|---|---|
| CritPt v2 (Hard) | Active | 1.5% | 93 | 2026-11-15 | independent | Source ↗ |
| EnterpriseOps-Gym-AA | Active | 18.4% | 93 | 2026-11-15 | independent | Source ↗ |
| AutomationBench-AA | Active | 0.8% | 93 | 2026-11-15 | independent | Source ↗ |
| MLCR-AA (Medical Long Context Reasoning) | Active | 2.2% | 93 | 2026-11-15 | independent | Source ↗ |
| GDP.pdf | Active | 3% | 93 | 2026-11-15 | independent | Source ↗ |
| FrontierCode-AA 2.0 | Active | 0% | 93 | 2026-11-15 | independent | Source ↗ |
| AA-Omniscience (Non-Hallucination) | Active | 64.3% | 93 | 2026-11-15 | independent | Source ↗ |
| AA-Omniscience (Accuracy) | Active | 14.7% | 93 | 2026-11-15 | independent | Source ↗ |
| HLE-Diamond | Active | 4.5% | 93 | 2026-11-15 | independent | Source ↗ |
| GPQA | Nearing Saturation | 74.3% | 78 | 2026-11-15 | independent | Source ↗ |
| CritPt | Active | 0% | 93 | 2026-11-15 | independent | Source ↗ |
| AA-LCR (Long Context Reasoning) | Active | 60.3% | 93 | 2026-11-15 | independent | Source ↗ |
| Terminal-Bench-Science 0.1 | Active | 0% | 93 | 2026-11-15 | independent | Source ↗ |
| Terminal-Bench 4.0 | Active | 0.5% | 93 | 2026-11-15 | independent | Source ↗ |
| Terminal-Bench 2.1 | Active | 24.3% | 93 | 2026-11-15 | independent | Source ↗ |
| τ³-Banking | Active | 8.9% | 93 | 2026-11-15 | independent | Source ↗ |
| SciCode | Active | 32.1% | 93 | 2026-11-15 | independent | Source ↗ |
| Humanity's Last Exam | Active | 10.6% | 93 | 2026-11-15 | independent | Source ↗ |
Local Hardware Execution (Ollama)Hardware Compatible
# 1. Pull and execute model locally with Ollama
ollama run nvidia-nemotron-3-5-lightning
# 2. Or invoke via local OpenAI-compatible endpoint
curl http://localhost:11434/v1/chat/completions \
-H "Content-Type: application/json" \
-d '{
"model": "nvidia-nemotron-3-5-lightning",
"messages": [{"role": "user", "content": "Analyze reasoning chains on AIME 2026"}]
}'Evaluation & Sourcing Notes
Scores listed for Nemotron 3.5 Lightning represent verified evaluations extracted from official research papers, independent evaluation suites (HELM, LMSYS, OpenCompass, LiveBench), and verified audit reports.
All benchmarks marked as saturated or deprecated reflect historical performance where the benchmark no longer provides active discriminative power.