Lab
Nvidia
Combined 3.97% · 30d -0.82pp
Nvidia publishes Nemotron models, tuned open-weights variants optimised for its own inference stack.
As of 2026-09-01, Tidelines tracks 2 Nvidia models on OpenRouter, together accounting for 3.97% of measured token share (-0.82pp over the last 30 days). Its most-used model is NVIDIA: Nemotron 3 Ultra at 3.04% of all tracked tokens, or 77% of Nvidia's own volume. The charts below break that down by model, tokens per day and estimated spend, updated daily.
Nvidia tokens per day & est. spend per day
Combined lab tokens × list price (30/70 blend). 90-day archive · list-rate estimate.
Nvidia's combined tokens grew ~28× over the last 90 days of archive to 643.0B/day.
Nvidia combined share of tokens
Sum of all Nvidia models' share on OpenRouter. 30d change: -0.82pp.
Model mix over time
Each Nvidia model as a share of the lab's combined tokens.
Where the tokens go vs where the dollars go
Each Nvidia model's share of the lab's tokens paired with its share of estimated spend (latest day, list-rate).
NVIDIA: Nemotron 3.5 Lightning takes 23% of Nvidia's tokens but 2% of its estimated spend.
Citable stat
As of 2026-09-01, Nvidia's combined share on OpenRouter is 3.97% (-0.82pp / 30d). Source: Tidelines.
Models
| Model | Share ↓ | Tokens/d | Est. $/d | 7d Δ | 30d Δ | $/M out |
|---|---|---|---|---|---|---|
| NVIDIA: Nemotron 3 Ultra | 3.04% | 493B | $1.17M | -1.16 | -1.56 | $3.13 |
| NVIDIA: Nemotron 3.5 Lightning | 0.93% | 150B | $25K | 0.00 | +0.74 | $0.20 |
as of 2026-09-01