Cerebras Inference vs Replicate pricing

Verified 2026-10-23 (2 days ago)Source: Cerebras Inference: www.cerebras.ai/pricing, Replicate: replicate.com/pricingVerified shows the older of the two.

Monthly cost at three volumes

Monthly cost of Cerebras Inference and Replicate at three volumes and at your own
Volume	Cerebras Inference	Replicate
small1M input, 200K output	$0.50 / moGPT OSS 120B	$5.75 / modeepseek-ai/deepseek-r1
growing50M input, 10M output	$25 / moGPT OSS 120B	$287.50 / modeepseek-ai/deepseek-r1
scale1B input, 200M output	$500 / moGPT OSS 120B	$5,750 / modeepseek-ai/deepseek-r1
your volumeset above	$25 / moGPT OSS 120B	$287.50 / modeepseek-ai/deepseek-r1

Each cell is the cheapest feasible plan, or the chosen model, at that volume, from the shared Overage cost function. "Cheapest model" only considers models with both an input and an output price listed. Volumes are example scenarios you can edit.

Cheapest model per service, USD per 1M tokens

Cheapest priced model of each service
Service	Cheapest model	Input	Output	Models listed
Cerebras Inference	GPT OSS 120B	$0.35	$0.75	2
Replicate	deepseek-ai/deepseek-r1	$3.75	$10	1

Every model

Every model of both services, USD per 1M tokens
Model	Service	Input	Output
GPT OSS 120B	Cerebras Inference	$0.35	$0.75
Qwen 3.8 27B	Cerebras Inference	$0.99	$1.49
deepseek-ai/deepseek-r1	Replicate	$3.75	$10

Models are compared by name and price, never by position. Pick one model per service, or "Cheapest model", in the cost table above.

Tiers, aligned by position

Tiers of Cerebras Inference and Replicate aligned by position: free, entry, mid, top
Position	Cerebras Inference	Replicate
Free	no tier	no tier
Entry	unlistedDeveloper*	no tier
Mid	no tier	no tier
Top	Contact salesEnterprise	no tier

Tightest rate limit

Tightest documented rate limit of each service
Measure	Cerebras Inference	Replicate
Tightest limit	unlisted	unlisted

Changes in the last 90 days

No changes recorded.

No change detected for either service in the last 90 days.