Cerebras ditches plan cards for per-model token pricing starting at $0.35/M

WHY IT MATTERS

For an inference provider whose core pitch is raw speed, abandoning packaged tiers for transparent per-token rates is a bet that price transparency, not plan structure, is what wins developer trust in a crowded GPU-inference market. Watch whether Cerebras keeps this table stable for more than a few weeks, or whether the same 10-day cadence produces another restructuring before it settles.

BY THE NUMBERS

• Cerebras's 3rd credits/usage-billing change since Jul 20, 2026
• 7 real changes in 90 days, ~1 every 10 days; this one just 5 days after the last
• GPT OSS 120B: $0.35/$0.75 per M tokens; Qwen 3.8 27B: $0.99/$1.49 per M tokens
• 596 pricing changes across 330 companies touched usage billing in 180 days (10.7%)

THEIR RATIONALE

Cerebras frames this as making its pricing transparent and usage-based rather than bundled into flat tiers, continuing a run of pricing-page iteration that has produced 7 real changes in the last 90 days at roughly one every 10 days. This lands just 5 days after Cerebras rewrote its Free Trial positioning around speed, suggesting the company is actively tuning both its message and its pricing structure in tandem rather than treating this as a one-off update.

THE SIGNAL

This is Cerebras's third move on credits and usage billing since mid-July: renaming Free to Free Trial with $5 credit framing, then a Developer Tier per-model rate table (where GPT OSS 120B briefly sold out at the same $0.35/$0.75 rate), and now a full retirement of plan cards in favor of straight per-token pricing. Usage-based credit billing is common right now — 596 pricing changes across 330 companies touched the pattern in the last 180 days — but Cerebras is cycling through it faster than almost anyone in that group, treating its pricing page as a live draft rather than a settled structure.

IN CONTEXT

Cerebras has touched credits and usage billing three times in under two months — renaming Free to Free Trial with $5 credit framing in July renaming Free to Free Trial, then revealing a per-model rate table a per-model rate table, and now retiring its plan cards for a straight Developer/Enterprise comparison. That's the fastest pace in the index: seven real changes in under two months, this one landing five days after Cerebras rewrote its Free Trial pitch rewrote its Free Trial pitch. It joins roughly 330 companies — including HubSpot and Datadog — that shifted pricing toward usage-based credits in the last six months, though few touch the pattern this often. For a speed-focused inference provider, ditching tiers for raw per-token rates bets that transparency, not packaging, is now the differentiator.