Epoch AI measures how fast inference prices are falling

Made by AIWritten by AI (Claude) from the source below. Check anything important against the source.How it's made

Measured the rate of decline in inference prices — 9× to 900× per year depending on the benchmark

Why it matters

The first quantitative measurement breaking down, task by task, the steep drop in token prices at equal performance

Source — Primary sources first; secondary reports are labelled “Press”

Officialepoch.aihttps://epoch.ai/data-insights/llm-inference-price-trends

Added in Week 35, 2026 · See that week

On the map — ⑧ Inference Infrastructure line

  1. Previous stopV1 engine in alpha
  2. This recordEpoch AI measures how fast inference prices are falling
  3. Next stopGemma 3 QAT int4
See it on the map

Same week

Figure announces BotQOfficialChina issues measures for labelling AI-generated and synthetic contentGovernmentCohere releases Command AOfficialAuditing Language Models for Hidden ObjectivesOfficial

Nearby on the same axis

NVIDIA announces Dynamo at GTCOfficialDeepSeek discloses the cost of its V3/R1 inference systemOfficialDeepSeek holds Open Source WeekOfficialGemma 3 QAT int4Official