VerdictPal
Join

Models / NVIDIA

Nemotron 3.5 Lightning

Nemotron · Released 2026-08-11

NVIDIA's 30B-class open Nemotron for local and single-GPU agent steps. Artificial Analysis Intelligence Index 23.6. AA hosted list $0.05 / $0.20 per 1M tokens.

Cheap, small Nemotron for routed agent steps, not a Spark or Opus replacement. Prefer sourced ledger rows over vendor speed claims.

#139 of 166 on AA Index · current snapshot
24 AA Index · 9 benchmark rows · 1 sourcesPosition is benchmark-specific — not a cross-family or cross-source ranking.
Composite33
ContextUnlisted
Input / 1M$0.05
Output / 1M$0.20
Knowledge cutoffunpublished
Output speed290 t/s
TTFT (default API)0.55s
Statussolid

Family profile

Best published score in each covered benchmark family.

46/100 avg
Reasoning
74
Coding
32
Agentic
24
Long context
55

4 tested benchmark families

Editor's note
Recorded from the 2026-08-15 AA snapshot (index 23.6). NVIDIA positions this as a distilled single-GPU agent specialist. Editorial depth still needs a desk pass.
Benchmark placements

Where Nemotron 3.5 Lightning places on each public benchmark source that publishes a row. Rank counts every model with a latest row in the same test — not a universal quality score.

BenchmarkFamilyRankScoreSourceDate
AA output speedPerformance#7/ 185Artificial Analysis2026-09-01
AA time to first tokenPerformance#76/ 185Artificial Analysis2026-09-01
τ³-Bench BankingAgentic#93/ 1069Artificial Analysis2026-09-01
AA-LCRLong context#146/ 17955Artificial Analysis2026-09-01
Humanity's Last ExamReasoning#154/ 18011Artificial Analysis2026-09-01
GPQA DiamondReasoning#157/ 18574Artificial Analysis2026-09-01
Artificial Analysis Intelligence IndexReasoning#158/ 18424Artificial Analysis2026-09-01
SciCodeCoding#162/ 18032Artificial Analysis2026-09-01
Terminal-BenchAgentic#163/ 18424Artificial Analysis2026-09-01
Family context

The three highest-scoring models with pages in each capability family. Where Nemotron 3.5 Lightning shows up, it's highlighted.

ReasoningNemotron 3.5 Lightning · 74
  1. 1
    Claude Mythos Preview
    95
  2. 2
    Gemini 3.7 Flash
    95
  3. 3
    Grok 4.6
    95
CodingNemotron 3.5 Lightning · 32
  1. 1
    GPT 5.5
    96
  2. 2
    Claude Opus 4.8
    96
  3. 3
    Claude Opus 4.6
    96
AgenticNemotron 3.5 Lightning · 24
  1. 1
    GPT 5.5
    100
  2. 2
    Claude Opus 4.8
    100
  3. 3
    Claude Fable 5
    100
Long contextNemotron 3.5 Lightning · 55
  1. 1
    Muse Spark 1.2
    83
  2. 2
    Kimi K3
    83
  3. 3
    Gemini 3.5 Flash
    81
Newest receipts

Benchmark rows added to the public ledger for Nemotron 3.5 Lightning in the last 120 days. Older rows live in the full table below.

9 recent
BenchmarkFamilyScoreSourceDays ago
AA-LCRLong context55.33%Artificial Analysis2
AA output speedPerformance290 t/sArtificial Analysis2
AA time to first tokenPerformance0.55sArtificial Analysis2
Artificial Analysis Intelligence IndexReasoning23.6%Artificial Analysis2
GPQA DiamondReasoning74.3%Artificial Analysis2
Humanity's Last ExamReasoning10.6%Artificial Analysis2
SciCodeCoding31.6%Artificial Analysis2
τ³-Bench BankingAgentic8.87%Artificial Analysis2
Terminal-BenchAgentic24.34%Artificial Analysis2
Family coverage

A high composite that hides a weak family is a trap. These bars surface the families where this model hasn't been publicly tested, and where it leads.

Reasoning74
Coding32
Agentic24
Long context55
Price vs. performance
0255075100$0.10$0.30$1$3$10$30input USD / 1M tokens · log scale →← intelligence

Hover or tab any dot for name, score, and input price.

AnthropicOpenAIMistralAlibabaDeepSeekGoogleMoonshotMetaMiniMaxxAIZhipuXiaomiOtherMeituanCohereNVIDIA

Full benchmark ledger

Every catalog benchmark for Nemotron 3.5 Lightning. Scores link to the original source; gaps mean no public row exists yet.

9 sourced rows

9 of 37 catalog benchmarks have a sourced row for Nemotron 3.5 Lightning.

Reasoning

Coding

Agentic

Long context

Performance

Changelog
  1. Nemotron 3.5 Lightning on Artificial Analysis (index 23.6, $0.05 / $0.20).
Sources