Skip to content
ModelCap

Nemotron 3.5 Lightning

Official publisher releaseRestricted licenseFree tier6.9d #220 → #219

NVIDIA·#219 of 371 on the current board·cheaper than #198 GPT-4o (2024-08-06)

Nemotron 3.5 Lightning is a language model from NVIDIA. It ranks #219 of 371 on ModelCap's current board, with an Index score of 34.9, modeled from the publisher's launch results against measured models. Its score range of 21.1–48.7 spans positions #143–#304. It is listed at $0.07 per million input tokens and $0.20 per million output tokens across 5 providers. It reads up to 1M tokens of context, and its weights are downloadable under a use-restricted licence.

Figures as of 19 Nov 2026, 05:16 UTC · download the public dataset · methodology

ModelCap Index · Current board
34.9#219
Modeled from the publisher's launch results against measured models · range 21.1–48.7
Market Gravity
57.1Increased by 0.3
Secondary market signal
Input
$0.07
per 1M tokens
Output
$0.20
per 1M tokens
Context
1M
tokens
Providers
5
endpoints

Providers

Uptime measured over the last 30 minutes
ProviderInputOutputContextQuantUptime
Darkbloomint4Cheapest$0.025$0.12262Kint4100.0%
DeepInfrabf16$0.06$0.16262Kbf16100.0%
CoreWeavebf16$0.07$0.20262Kbf16100.0%
Fireworks$0.05$0.20262K—100.0%
ModelRunfp4$0.07$0.201Mfp4100.0%

Hugging Face Inference Providers

Exact repository identity

Provider economics and telemetry published by the Hugging Face Router. These are operational signals, not benchmark results.

ProviderInputOutputContextTTFTThroughputFeatures
deepinfra
live
$0.06$0.16262K279 ms175.0 tok/stools · structured
coreweave
live
$0.07$0.20262K238 ms178.0 tok/stools · structured
fireworks-ai
live
$0.05$0.20262K273 ms444.5 tok/stools · structured
modelrun
live
$0.07$0.201M114 ms262.0 tok/stools · structured

Reported evaluation leads (7)

Not independent measurement

Structured rows reported in this exact model repository at the pinned revision below. They are the publisher talking about itself: never measured evidence, never a score input on their own. Rows marked as used were laddered against independently measured peers inside the launch-card channel of this model's modeled placement, with the replay-measured optimism haircut applied; the others played no part.

Dataset / taskMetricValueRevisionProvenance
allenai/IFBench_test
prompt_level_loose
—71.88
Unpinned
not published
Reported
Not used in launch estimate
cais/hle
hle
—11.72
Unpinned
not published
Reported
Not used in launch estimate
harborframework/terminal-bench-2.1
terminalbench_2_1
—24.58
Unpinned
not published
Reported
Used in launch estimate
Idavidrein/gpqa
diamond
—75.44
Unpinned
not published
Reported
Not used in launch estimate
SWE-bench/SWE-bench_Multilingual
swe_bench_multilingual_%_resolved
—39.33
Unpinned
not published
Reported
Used in launch estimate
SWE-bench/SWE-bench_Verified
swe_bench_%_resolved
—51.56
Unpinned
not published
Reported
Used in launch estimate
TIGER-Lab/MMLU-Pro
mmlu_pro
—81.94
Unpinned
not published
Reported
Used in launch estimate

ModelCap Index · Current board

Methodology
34.9
Modeled from the publisher's launch results against measured models
Current board
#219 of 371
Score range
21.1–48.7
Rank range
#143–#304
Technical details
Public rank basis
Modeled · peer benchmarks (architecture)
Evidence label
Modeled · peer benchmarks · publisher-corpus-prior over 214 held-out anchors (22%); launch card against 3 resolved peers on 11 rows (78%); optimism haircut 2.8 from cross-lab-probe; shrunk 0.5 up toward the measured corpus
Index support
38.9% · limited
Identity binding
Exact catalogue product · nvidia/nvidia/nemotron-3.5-lightning/bf16 · aggregates disclosed benchmark configurations · endpoint configuration not claimed · artifact metadata nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16@a9904d24bcc1d289a1950fa9d2b978c47cf903b9 (not the evaluation revision)
Index observations used
0
Qualified Index evidence point
34.9
How this position was produced · fused launch placement

No general-family board lists this model yet. Its point estimate is the precision-weighted mean of every channel that spoke for it; each channel’s share is the weight it carried. Independent measurements replace this placement as boards list the model.

Fused estimate
34.9 ± 10.8
Independent share
0%
Shrink toward corpus
+0.5
Peer ceiling share
0%
Optimism haircut
−2.8 points · cross lab probe, 3 probes
Method
precision-weighted-v1
Channels fused into this model’s launch placement
ChannelEstimateShare
Publisher corpus prior · not independent
publisher-corpus-prior over 214 held-out anchors
36.6 ± 22.922%
Launch card against resolved peers · not independent
launch card against 3 resolved peers on 11 rows
34.4 ± 12.278%
  • · Launch card row aa-lcr: inside the resolved peer range
  • · Launch card row aa-omniscience: inside the resolved peer range
  • · Launch card row gpqa-diamond-without-tools: inside the resolved peer range
  • · Launch card row hle-text-only-without-tools: claims below every resolved peer
  • · Launch card row ifbench: inside the resolved peer range
  • · Launch card row mmlu-pro: inside the resolved peer range
  • · Launch card row scicode: claims below every resolved peer
  • · Launch card row swe-bench-multilingual: claims below every resolved peer
  • · Launch card row swe-bench-verified: claims below every resolved peer
  • · Launch card row tau-3-bench: claims below every resolved peer
  • · Launch card row terminal-bench-2-1: inside the resolved peer range
Bayesian Evidence Score · measured evidence input only

BES normalizes admitted public benchmark evidence for measured Index rows. Its legacy score and observables prior do not define the public language rank.

Observed capability
—
Evidence status
No admitted measured evidence
General preference
—
Coding
—
Agents & tools
—
Reasoning
—
Evidence breadth
0%
Evidence coverage
0%
Market and catalogue signals · secondary, never the language Index rank
Usage (OpenRouter popularity)
42.4
Liquidity (providers × uptime)
91.7
Open reach (HF downloads)
—
Surface (context / tools / modalities)
85.0
Freshness
68.4
  • · OpenRouter popularity #29
  • · 5 live providers
  • · Catalogue: context 1048576, reasoning, tools
  • · First seen 2026-08-11

These adoption and deployment observations remain context only. They do not change this model's ModelCap Index score or rank.

The bounded 0–100 capability score blends each source’s competitive placement with its published achievement, then combines capability families. Evidence breadth adds a modest uncertainty adjustment; price and popularity are not part of benchmark-led rank.

Experimental capability estimate

Estimate policy
55.4
41.6–69.1 uncertainty interval
Excluded from rank
Confidence
78%
Training support
225 models · 132 lineages
Out-of-domain check
in domain
Feature coverage
100%
Lineage-held-out error
8.1 MAE

Experimental estimate. This layer predicts from admitted evidence and safe metadata only. It never enters ModelCap Score or rank.

Deployment readiness

Metadata index
74.0
partial · 76% metadata coverage
Not capability
Artifact reproducibility
85.0
Access & legal clarity
92.5
Deployability
100.0
Evaluation provenance
0.0

Not yet published: the base model it derives from, admitted benchmark results, results across several capability areas, results from more than one source, a confident match to its benchmark entries, pinned evaluation dataset versions.

A metadata completeness and deployability index—not a safety certification, quality grade, or production-readiness claim.

Market activity

Venue coverage
Market Gravity
57.1
OpenRouter weekly popularity
#29 of 345

Market Gravity uses OpenRouter’s full-catalogue popularity order. Sparse Vercel top-ten observations appear only when published and do not affect the score.

Arena preference

No matched Arena rating

This model is visible because it is available in the live catalogue, but it has no trustworthy community result yet. ModelCap does not infer a quality score from price, features, or market activity.

Market Gravity breakdown

Usage55%
42.4

OpenRouter weekly popularity across the full model catalogue.

Liquidity25%
91.7

Independent providers versus the model's open or closed cohort, weighted by uptime.

Open reach15%
50.0

Hugging Face 30-day downloads for open models; neutral for closed models.

Freshness5%
68.4

Time since first public availability, on a six-month half-life.

Specification

Model ID
nvidia/nemotron-3.5-lightning
Context window
1M tokens
Declared output cap
236K tokens

OpenRouter catalogue or provider-endpoint declaration. Limits can vary by provider and prompt length; a model-wide maximum has not been independently verified. OpenRouter metadata

Inputs
Text
Outputs
Text
Tokenizer
Other
Cached input tokens
$0.04 / 1M
First seen on OpenRouter
11 Aug 2026

Capabilities

  • Supported: Reasoning
  • Supported: Tool use
  • Supported: Structured output
  • Supported: Response format
  • Not supported: Moderated

Weights & access

Restricted license. Downloadable weights use a model-specific or use-restricted license. Review the linked license before use, redistribution, or commercial deployment.

Access
Restricted license
Pinned revision
a9904d24bcc1
Parameters
31.6B
Architecture
NemotronHForCausalLM
Model type
nemotron_h
Hugging Face downloads (30d)
974K
Downloads (all time)
2.4M
Likes
278
Repository updated
24 Aug 2026, 23:46 UTC

Benchmark leaderboards featuring Nemotron 3.5 Lightning

All leaderboards
Public benchmark boards on which Nemotron 3.5 Lightning has a published result
LeaderboardScoreSource rankConfigurationPublished
Artificial Analysis Intelligence IndexArtificial Analysis5.5#114 of 125Reasoning29 Oct 2026

Alternatives to Nemotron 3.5 Lightning, by the numbers

Ranked neighbours

Models within 3 places of #219 on the ModelCap Index.

Cheaper at a similar or higher Index

Index score within 3 points of, or above, this model's, with a lower listed blended price (3:1 input:output).

Other ranked NVIDIA models

The same publisher's other current models with a public Index position.

Nemotron 3.5 Lightning: common questions

How does Nemotron 3.5 Lightning rank among AI models?

Nemotron 3.5 Lightning holds ModelCap Index position #219 of 371 ranked language models as of 19 November 2026, with an Index score of 34.9 (interval 21.1–48.7); evidence: modeled · peer benchmarks. The Index combines public benchmark boards with published uncertainty; the methodology page explains the weighting.

How much does Nemotron 3.5 Lightning cost per 1M tokens?

Nemotron 3.5 Lightning is listed at $0.07 per 1M input tokens and $0.20 per 1M output tokens as of 19 November 2026; the lowest output price among 5 providers is $0.12 on Darkbloom. Prices come from the OpenRouter catalogue and re-observe on every refresh; at least one provider lists a free tier. For 1,000 requests using 2,000 input and 500 output tokens each, the listed rates imply $0.24 total (2M input + 0.5M output). This excludes caching, batch discounts, prompt-length tiers, tool charges and retries; verify the chosen endpoint's rate and limits before budgeting.

What is the context window of Nemotron 3.5 Lightning?

Nemotron 3.5 Lightning has a published context window of 1M tokens (1,048,576). The snapshot lists a 235,929 token catalogue-declared route output cap. OpenRouter catalogue or provider-endpoint declaration. Limits can vary by provider and prompt length; a model-wide maximum has not been independently verified. The providers table lists each endpoint's context limit.

Is Nemotron 3.5 Lightning open-weight?

Restricted license (openmdw-1.1). Downloadable weights use a model-specific or use-restricted license. Review the linked license before use, redistribution, or commercial deployment. The repository is nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16 on Hugging Face.

Which API providers serve Nemotron 3.5 Lightning?

5 providers serve Nemotron 3.5 Lightning through OpenRouter as of 19 November 2026: Darkbloom, Fireworks, DeepInfra, ModelRun and CoreWeave. Each provider's price, context limit, quantization and measured uptime are in the providers table above.

Which benchmarks has Nemotron 3.5 Lightning been evaluated on?

Nemotron 3.5 Lightning has published results on 1 public board tracked by ModelCap as of 19 November 2026: AA Intelligence Index 5.5 (#114 of 125). Each board page ranks every tracked model on that benchmark. The ModelCap Index scores the results its methodology admits, with published uncertainty, and shows reference boards without scoring them.

When was Nemotron 3.5 Lightning released?

Nemotron 3.5 Lightning first appeared in the catalogue on 11 Aug 2026. It is the current version in its family. Rank and price movements since then are recorded on the site's changes feed.