LLM Opex

Changes

DeepInfra cut nvidia/NVIDIA-Nemotron-3-Super-120B-A12B (1M in), priority: $0.1275 → $0.1125/1M in (-12%)

DeepInfra · recorded 3 November 2026

Before: $0.1275 per 1M input tokens. Now: $0.1125 per 1M input tokens (-11.8%).

What it means: At 6M input tokens a day this meter costs $20.25 a month, was $22.95.

This price over the stored collections

DeepInfra prices · November 2026 report