Find the price behind the plan

Search by provider, model, hardware, or billing trap. Use this when you want the official source before choosing API, subscription, GPU-hosted, or hybrid architecture.

Browse by price type

64 matching prices

Official prices

Showing 17 of 64 matching prices.

Only published provider prices. No guessed prices. 2 official pricing sources need review.

ProviderPrice
AnthropicToken

Claude Haiku 4.5

$1/$5 per 1M

2026-09-03

output-token multiple, regional endpoint premium
AnthropicSub

Claude Max 20x

$200/mo

2026-09-03

web subscription only, mobile pricing may vary
AnthropicSub

Claude Max 5x

$100/mo

2026-09-03

web subscription only, mobile pricing may vary
AnthropicToken

Claude Opus 4.8

$5/$25 per 1M

2026-09-03

premium reasoning tier, high output-token multiple
AnthropicToken

Claude Sonnet 5

$2/$10 per 1M

2026-09-03

promo windows can expire, high output-token multiple
AWS BedrockToken

Claude Sonnet on Bedrock

$2/$10 per 1M

2026-08-10

region and marketplace prices vary, provisioned throughput is separate
AWS EC2GPU

p5.48xlarge Capacity Block H100

$5.19/GPU hr

2026-09-03

capacity-block reservation required, region availability varies
Azure AI Foundry ModelsToken

Azure AI Foundry token price

$0.125/$0.5 per 1M

2026-09-03

region and deployment option can change price, provisioned throughput is separate
Google Vertex AIToken

Gemini 2.5 Flash on Vertex AI

$0.3/$1 per 1M

2026-09-03

global vs regional pricing can differ, batch and provisioned throughput are separate
GroqToken

GPT OSS 20B

$0.075/$0.3 per 1M

2026-08-04

quality ceiling, tool calls add separate charges
OpenAISub

ChatGPT Plus with Codex

$20/mo

2026-07-03

usage limits apply, not a production API substitute
OpenAIToken

GPT-5.5

$4/$5 per 1M

2026-09-03

high output-token multiple, priority tier premium
OpenAIToken

o4-mini fine-tuning

$100/training hr

2026-09-03

training failures still cost money, new-user access restricted
ReplicateCredit

Prepaid credit auto reload

$5 threshold / $15 reload

2026-09-03

cash locked before usage, zero balance stops new work
Vast.aiGPU

H100 marketplace reference

$1.47/GPU hr

2026-09-03

host reliability varies, bandwidth and storage can change total cost
Vast.aiEgress

Internet upload bandwidth

$0.0000/GB egress

2026-09-03

marketplace host bandwidth varies, egress can dominate dataset export jobs
Vercelpass through

AI Gateway provider pass-through

provider list price

2026-09-03

provider price still applies, credit balance can block fallback

Cost shortcuts

Jump to the row you probably need first.

Each shortcut filters the official table above to one source-backed price.

Price checks

Source, route, trap, freshness.

Use these checks after you find a row, before moving work to a provider.

Can I verify it?

Yes

Every visible price links to a provider source and observation date.

Can I compare routes?

Yes

Prices cover API tokens, subscriptions, GPU rental, cloud AI, routing, tuning, storage, egress, and credits.

Can I spot hidden costs?

Yes

The table calls out output-token multiples, cache misses, idle GPU time, egress, storage, and prepaid credit rules.

Is anything stale?

Review

Some provider pages need review before new prices appear.

How Price Gouge keeps this table usefulOfficial links, dates, and stale-price warnings

Results are grouped around buyer decisions: API tokens, subscriptions, GPU hosting, tuning, storage, egress, and credits.

Every visible price keeps its provider link and observation date so you can verify the number before committing spend.

Some provider pages need review before new prices appear. 33 of 35 official sources are current. Latest verified source update: Sep 3, 2026, 8:18 AM UTC. Next automatic source review: Sep 4, 2026, 8:17 AM UTC.