BenchLeader

NVIDIA vs Thinking Machines

Verdict
  • Thinking Machines's flagship leads: Inkling 2 Small at 57.9 vs Nemotron 3 Ultra 550B A55B at 55.4.
  • Thinking Machines has the better model in 4 of 4 price bands.
  • NVIDIA's fastest ranked model (Nemotron 3.5 Super VL, 271 tok/s) beats Thinking Machines's (Inkling Small, 116 tok/s).
  • NVIDIA publishes more open-weights models (21 vs 2).

Best model per price band

Each provider’s highest-index model below the price ceiling. Purple marks the winner of the band.

Price bandNVIDIAThinking Machines
Under $0.50/MNemotron 3 Super 120B A12B thinking 51.1$0.450/MInkling 2 Small 57.9$0.475/M
Under $2/MNemotron 3 Ultra 550B A55B 55.4$1.20/MInkling 2 Small 57.9$0.475/M
Under $10/MNemotron 3 Ultra 550B A55B 55.4$1.20/MInkling 2 Small 57.9$0.475/M
Any priceNemotron 3 Ultra 550B A55B 55.4$1.20/MInkling 2 Small 57.9$0.475/M

Line-up

MetricNVIDIAThinking Machines
Flagship index55.457.9
Ranked models113
Open-weights models212
Cheapest ranked, $/M$0.088$0.475
Fastest ranked, tok/s271116
Release cadence, days36–
Last release23 Oct 2026–
Next expected13 Nov 2026 – 12 Dec 2026–

Data as of 16 Nov 2026. See NVIDIA and Thinking Machines in full.