LLM Pricing
ModelsLeaderboardsToolsProvidersReleasesArticles

© 2026 LLM Pricing

About
·Contact
·Privacy
·RSS
← Model list

Nemotron 3.5 Lightning 30B A3B

nvidia·nvidia/nemotron-3.5-lightning·GA·Open weights·nemotron series
Fast NVIDIA Nemotron MoE for reliable agentic tasks across enterprise workloads

Specs & pricing

Input / output per 1M tokens
Official price·Nvidia
$0 / $0
Blended $0 · Cache read —
Lowest paid·Kilo GatewayGateway
$0.025 / $0.12
Blended $0.049 · 7.4× spread
9 more $0 channels
Context
262,144
Output limit
262,144
Knowledge cutoff
2025-09
Released / updated
2026-08-11 / 2026-08-11
Capabilities
✓ Reasoning✓ Tool use✓ Structured output✓ TemperatureAttachments
⚠ Providers report capability flags inconsistently
Modalities
Text

Nemotron 3.5 Lightning 30B A3B by nvidia is offered by 21 providers on this page. Public prices are shown for 12 of them. Its official list price is $0 per 1M tokens. No paid channel is priced below the official list price. Prices vary about 7.4× across the providers that host it. It also has 9 free ($0) channels; free tiers usually carry rate limits, and subscription-covered access bills $0 per token only after the subscription fee.

Artificial Analysis rates it 5.5 on the Intelligence Index, with 15.5 for Engineering, its strongest domain. Against its lowest blended price of $0.049 per 1M, that is roughly 113 index points per dollar, which is the value ratio the leaderboards rank on. Median output speed is 297 tokens per second, with 0.63s to the first token. Running one task of the Artificial Analysis suite costs about $0.3145, which reflects how many tokens its reasoning consumes rather than the unit price alone.

The context window is 262,144 tokens at the reference host, but hosts report different limits, from 65,536 to 1,048,576, so the usable window depends on the provider you pick. It supports reasoning, tool use, and structured output. The weights are open, so it can also be self-hosted or served through a gateway of your choice. Providers report the capability flags inconsistently, so verify a specific feature against the host you plan to use. Its training knowledge cuts off at 2025-09.

Quality & performanceArtificial Analysis · Intelligence Index v5

Intelligence5.5
Value113
Engineering15.5
Finance & Accounting14.9
Legal14.2
Healthcare & Medical9
Strategy & Ops7.1
Economics14.8
Output speed297 tok/s
TTFT0.63 s
Cost per task$0.31
Value formulaIQ 5.5 ÷ min blended $0.049 = 113

Quality is independently evaluated by Artificial Analysis. Speed/latency are model-level medians.

Available at 21 providers12 with public prices · 9 free

  • ProviderInput / Output
  • NvidiaOfficialCloud
    nvidia/nemotron-3.5-lightning-30b-a3b
    Free
    Context 262,144Output limit 262,144
  • RequestyGateway
    nemotron-3.5-lightning-30b-a3b
    Free
    Context 1,048,576 ⚠Output limit 65,536
  • Merge GatewayGateway
    nvidia/nemotron-3.5-lightning-30b-a3b
    Free
    Context 1,000,000 ⚠Output limit 262,144
  • OpenRouterGateway
    nvidia/nemotron-3.5-lightning:free
    Free
    Context 1,000,000 ⚠Output limit 65,536
  • AIHubMixGateway
    nemotron-3.5-lightning-free
    Free
    Context 262,144Output limit 262,144
  • Kilo GatewayGateway
    nvidia/nemotron-3.5-lightning:free
    Free
    Context 1,000,000 ⚠Output limit 65,536
  • InferAllGateway
    nvidia/nemotron-3.5-lightning-30b-a3b
    Free
    Context 262,144Output limit 262,144
  • Vercel AI GatewayCloud
    nvidia/nemotron-3.5-lightning-free
    Free
    Context 1,000,000 ⚠Output limit 65,000
  • OpenCodeGateway
    nemotron-3.5-lightning-free
    Free
    Cache read $0Context 262,144Output limit 262,144
  • Kilo GatewayGateway
    $0.025 / $0.12
    Cache read $0.013Context 262,144Output limit 235,929
ProviderTierInputOutputCache readCache writeContextOutput limitStatus
NvidiaOfficial
nvidia/nemotron-3.5-lightning-30b-a3b
CloudFree——262,144262,144
Requesty
nemotron-3.5-lightning-30b-a3b
GatewayFree——1,048,576 ⚠65,536
Merge Gateway
nvidia/nemotron-3.5-lightning-30b-a3b
GatewayFree——1,000,000 ⚠262,144
OpenRouter
nvidia/nemotron-3.5-lightning:free
GatewayFree——1,000,000 ⚠65,536
AIHubMix
nemotron-3.5-lightning-free
GatewayFree——262,144262,144
Kilo Gateway
nvidia/nemotron-3.5-lightning:free
GatewayFree——1,000,000 ⚠65,536
InferAll
nvidia/nemotron-3.5-lightning-30b-a3b
GatewayFree——262,144262,144
Vercel AI Gateway
nvidia/nemotron-3.5-lightning-free
CloudFree——1,000,000 ⚠65,000
OpenCode
nemotron-3.5-lightning-free
GatewayFree$0—262,144262,144
Kilo GatewayGateway$0.025$0.12$0.013—262,144235,929

Sorted by blended price (input×0.75 + output×0.25) asc. The official channel always shows regardless of rank. Whether a gateway's low price is actually usable can't be verified.

Reasoning control

Toggle (on / off)effort = noneeffort = loweffort = mediumeffort = higheffort = maxbudget_tokenseffort = xhigh

Interleaved thinking (reasoning between tool calls) is declared by 4 of 21 providers.

1 / 21 providers expose no reasoning control (reasoning_options: []).

Your usage cost

Common scenarios
1Kilo Gateway$9.50
2RunInfra$12.70
3NanoGPT$15.20
4NanoGPT$15.20
5Requesty$15.20
6Vercel AI Gateway$16.40

9 more channels offer $0 (Nvidia, Requesty, Merge Gateway etc.); free tiers usually have rate limits and no SLA, excluded from ranking.

The cheapest paid channel is the only channel.

Price historyone sample accumulated per data sync

Input listOutput listMin blended
$0.088$02026-08-122026-10-09
2026-10-09Input list $0Output list $0Min blended $0.049

Benchmark14 items

NameConditionsScoreMetricSource
MMLU-Proharness: NeMo Gym / NeMo Evaluator SDK · variant: BF16; reasoning81.94scoreSource ↗
AA-Omniscienceharness: NeMo Gym / NeMo Evaluator SDK · variant: BF16; reasoning17.5scoreSource ↗
SciCodeharness: NeMo Gym / NeMo Evaluator SDK · variant: BF16; reasoning32.6scoreSource ↗
PinchBenchharness: NeMo Gym / NeMo Evaluator SDK · variant: BF16; reasoning85.37scoreSource ↗
BrowseCompharness: NeMo Gym / NeMo Evaluator SDK · variant: BF16; reasoning36.97scoreSource ↗
AA-LCRharness: NeMo Gym / NeMo Evaluator SDK · variant: BF16; reasoning52scoreSource ↗
GPQA Diamondvariant: BF16; reasoning; no tools75.44accuracySource ↗
Humanity's Last Examvariant: BF16; reasoning; no tools · dataset: text-only11.72accuracySource ↗
SWE-Bench Verifiedharness: NeMo Evaluator · variant: BF16; reasoning51.56accuracySource ↗
SWE-Bench Multilingualharness: NeMo Evaluator · variant: BF16; reasoning39.33accuracySource ↗
Terminal-Benchharness: NeMo Evaluator · variant: BF16; reasoning · v2.124.58accuracySource ↗
tau3-benchvariant: BF16; reasoning; banking9.28scoreSource ↗
GDPval-AAvariant: BF16; reasoning · v2832EloSource ↗
IFBenchvariant: BF16; reasoning; loose71.88accuracySource ↗

The same benchmark scores very differently across harness / dataset, so the qualifying conditions must be shown together.

Artificial Analysis evaluations21 items

GPQA Diamond74.3%
Humanity's Last Exam10.6%
HLE-Diamond4.5%
CritPt0.0%
CritPt v2 Hard1.5%
SciCode32.1%
FrontierCode 2.00.0%
Terminal-Bench 2.124.3%
Terminal-Bench 4.00.5%
Terminal-Bench-Science0.0%
τ³-Bench Banking8.9%
AutomationBench-AA0.8%
EnterpriseOps-Gym-AA18.4%
AA-Briefcase507
GDPval-AA646
GDP.pdf3.0%
AA-LCR60.3%
MLCR-AA2.2%
AA-Omniscience Index-15.7
AA-Omniscience Accuracy14.7%
AA-Omniscience Non-Hallucination64.3%

Individual evaluations run by Artificial Analysis, on the same reasoning tier as the intelligence score above. Each benchmark has its own task set and scale (percent correct, points or an index), so rows are not comparable with one another.

Related models

Nemotron 3.5 Super VL 120B A12Bsame series$0 / $0Nemotron 3.5 Super 120B A12Bsame series—nemotron-3-ultra-nvfp4same series—Nemotron 3.5 Content Safetysame lab$0 / $0
Data partly from models.dev (MIT) · Nvidia official docs ↗