Breaking

OpenAI, Anthropic Face Mounting Pressure to Cut AI Prices

June 15, 2026 at 15:29 EDT

Foundation Models
Sales & Business
AI Agents

OpenAI, Anthropic and Google are being forced into successive price cuts and overhauls of their pricing structures. Having long sustained premium prices through performance differentiation, the leading labs are confronting the limits of their pricing power amid intensifying competition and rising user cost sensitivity.

The Economics of Inference · AI Pricing Power

When Every Query Costs Money, Flat Subscriptions Break

AI's near-zero marginal cost assumption doesn't hold: each user query burns GPU and power. Soaring inference bills are pushing OpenAI, Anthropic and GitHub from flat plans toward usage-based pricing.

$8.67B

OpenAI inference spend, first 9 months of 2025 — about 2× its revenue

90%

Discount cached input can unlock as providers slice pricing finer

Of organizations see clear revenue or cost gains — though 88% use AI

Anthropic Inference Gross Margin

From deeply negative to healthy — the turnaround behind the pricing shift

−94% to −109%

Earlier
(loss-making)

38%

A year ago

~70%

By 2026

Pricing Sliced by Capability

Output price per 1M tokens — flagships stay premium, mini models serve light tasks

GPT-5.5

$30.00

Claude Opus 4.8

$25.00

GPT-5.4

$15.00

GPT-5.4-mini

$4.50

Some welcome it

More choice from price competition
Cost cuts via caching and batching
Shift light work to mini/nano or cheaper models

Developers push back

"Budgets are now unpredictable"
Credits run out fast under heavy use ($500–$2,000/mo)
Abrupt end of flat plans called a "rug pull"

With inference costs that can outpace revenue still baked in, the squeeze isn't over.

Expect continued pricing revisions, finer capability tiers, and harder choices about which model handles which task.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.