Get started
Together AI · Language models

NVIDIA Nemotron 3.5 Super VL

Every price this sku has held since Vitna first read it, and the document behind each one. 3 periods across 3 dimensions.

ProviderTogether AI
ClassLanguage models
Marketopen-weight serving
ProductNVIDIA Nemotron 3.5 Super VL
WeightsNVIDIA
Tierserverless
01 / 04Current
confirmed 22 Nov 14:08 UTC
NVIDIA Nemotron 3.5 Super VLTogether AI · language modelsUSD per 1M tokens

input $0.15 · cached input $0.02 · output $0.50

USD per 1M tokens. Confirmed 22 Nov 14:08 UTC.

Confirmed22 Nov 14:08 UTC
Citeciteincachedout
02 / 04Price history
3 meters1 period each

Every meter has held one price

The ledgerfirst read 2 Nov 2026 · last read 22 Nov 14:08 UTC
MeterPeriodPriceStateDocumentsSigned day
Output priceUSD per 1M tokens2 Nov 2026 → open$0.50current · confirmedopened f19bba92c79clast read b0002d1d6ed3together-pricing ↗citenot yet signed (2026-11-22)
Cached input priceUSD per 1M tokens2 Nov 2026 → open$0.02current · confirmedopened f19bba92c79clast read b0002d1d6ed3together-pricing ↗citenot yet signed (2026-11-22)
Input priceUSD per 1M tokens2 Nov 2026 → open$0.15current · confirmedopened f19bba92c79clast read b0002d1d6ed3together-pricing ↗citenot yet signed (2026-11-22)
03 / 04Third-party evidence
145 models with an index · 132 on a design arena board

Measured quality

Intelligence Index13.6Position 43 of 58, above 26% of the models scored.the intelligence leaderboard
Coding Indexnot scoredArtificial Analysis has not scored this model on its Coding Index.the coding leaderboard
Agentic Indexnot scoredArtificial Analysis has not scored this model on its Agentic Index.the agentic leaderboard
  1. pdf · Agents984 Elowin rate 22.6% · source rank 36the board, 36 ranked
  2. 3D scenes · Models1116 Elowin rate 30.3% · source rank 86the board, 113 ranked
  3. Code · Models1138 Elowin rate 32.3% · source rank 94the board, 120 ranked
  4. Data visualisation · Models1087 Elowin rate 26.4% · source rank 105the board, 110 ranked
  5. Game development · Models1082 Elowin rate 27.5% · source rank 107the board, 120 ranked
  6. UI components · Models1155 Elowin rate 33.1% · source rank 80the board, 116 ranked
  7. Websites · Models1157 Elowin rate 34.6% · source rank 92the board, 126 ranked
04 / 04The public list
retrieved 17 Nov 2026 · sha 0bb53a77713f

On OpenRouter

Context1.05M
Modalitiestext, image, video → text
Listed

input $0.15 · output $0.50 · cached input $0.02

Per million tokens, as OpenRouter listed it when the catalog was read. OpenRouter is an aggregator: this is its routed rate, not a provider's own pricing page. Retrieved 17 Nov 16:29 UTC.

openrouter.ai

input $0.15 · output $0.50

Read from www.together.ai, the provider's own page, and kept as a dated record. Confirmed 22 Nov 14:08 UTC. The same weights, served by one provider at its own list price.

the record for NVIDIA Nemotron 3.5 Super VL
Sources

Source: OpenRouter (openrouter.ai/models), as of 17 Nov 16:29 UTC. sha256 0bb53a77713f over 853,270 bytes.
Catalog: OpenRouter's public model list.
Intelligence, coding and agentic indexes: Artificial Analysis, as published by OpenRouter.
Elo, win rate and rank: Design Arena, as published by OpenRouter.
Vitna ran none of these evaluations and adjusted none of these figures.

How to check this

Every stretch above names the sha256 of the document it was read from. The record at /api/v1/evidence/{sha256} gives the URL, the fetch time and the agent that read it, which is what lets anyone fetch the page themselves and compare hashes. A period is confirmed when a second read at least an hour later agreed with it; one that reverted before that is marked a flap and kept, because nothing here is deleted. A root is cut once a day, after the day closes, so the newest reads sit in a day that has not been signed yet and say so rather than reading as a gap. The newest signed day here is 2026-11-21. Every root, its statement, its signature and the key that made it are at /api/v1/attestations. Every figure above carries a cite mark, which leads to a permanent dated citation of it in the three shapes people paste: a footnote, a short reference and a schedule line. Cite a figure is what those are and how long the urls keep answering. The two catalog sections above cite OpenRouter's public model list the same way, by the sha256 of the document read, and their figures are third parties' rather than the record's.