Tier
Major
Headquarters
Santa Clara, CA
Founded
1993
Flagship
DGX / Megatron / Omniverse
Models tracked
18
Artificial Analysis
Top intelligence
37.8
Artificial Analysis Index
Cheapest model
$0.00
blended /1M
Fastest output
323
tokens / sec
Flagship Model
NVIDIA AI’s top-scoring model on the Artificial Analysis Intelligence Index.
Most intelligent model
Nemotron 3 Ultra 550B A55B (Reasoning)
Released 2026-06-04
Intelligence
37.8
Coding
49.3
Math
—
In $/1M
$0.68
Out $/1M
$2.67
Speed t/s
199
Benchmark breakdown
Models & benchmarks
Every NVIDIA AI model on file — intelligence, coding, price, and speed.
Trajectory
NVIDIA AI’s model releases over time, by intelligence.
House strengths
Where NVIDIA AI’s lineup over-indexes, averaged across its models.
Capability profile
Best value
NVIDIA Nemotron 3 Nano 30B A3B (Reasoning)
Most intelligence per dollar — 14.2 index at $0.09/1M blended.
161 index points per $1
Governance
NVIDIA AI’s license posture and openness across its lineup.
Open weight
8
of 8 models
Proprietary
0
of 8 models
Commercial-OK
4
license permits
Modalities
Text
Via LLM Stats
Media models
Image, video, and speech models, ranked by arena Elo.
| Model | Type | Released | Elo | Rank |
|---|---|---|---|---|
| Cosmos3-Super-Image2Video-4Step | image to video | 2026-07 | 1265 | #17 |
| Cosmos3-Super-Image2Video | image to video | 2026-05 | 1250 | #26 |
| Cosmos3-Super-Text2Image (agentic) | text to image | 2026-05 | 1218 | #10 |
| Cosmos3-Super-Text2Image-4Step | text to image | 2026-07 | 1181 | #24 |
| Magpie-Multilingual 357M (Feb 2026) | text to speech | — | 1062 | #44 |
| Magpie-Multilingual 357M | text to speech | — | 1003 | #70 |
| Magpie Multilingual | text to speech | — | 938 | #78 |
| Sana Sprint 1.6B | text to image | 2025-03 | 935 | #131 |