AI pricing pages

Claude sells four individual plans, Free, Pro at $17 a month on annual, Max at $100 and Ultra at $200, and what separates them is usage itself. Max is 5x Pro in both the five-hour and weekly windows, Ultra is 10x Pro weekly with no 5-hour limit for verified accounts, and both top plans carry a monthly API credit. One URL carries individual, team and API pricing, and the FAQ prints the enterprise rate.

AssemblyAI bills speech-to-text by the hour of audio, the unit a transcription team already uses, and the default view shows two rates: Universal-3.5 Pro at $0.21 /hr and Universal-2 at $0.15 /hr. A selector across eight APIs, grouped under TRANSCRIBE, UNDERSTAND and INTERACT, swaps the rate table, so each product is priced in its own unit, from audio hours to LLM tokens to input characters. The FAQ carries the billing mechanics in plain words, including the new Text-to-Speech API at $15 per 1M characters.

Baseten layers a three-tier service ladder over published rate sheets. Basic is '$0 per month, pay as you go', and Pro and Enterprise carry 'Volume discounts available' with 'GET A QUOTE'. Underneath sit three meters: Model APIs per 1M tokens across 19 model rows with Input, Cache Input and Output columns, opening with joey-1 at $0.01 input, dedicated GPU and CPU instances billed 'down to the minute', and the same GPU grid for training.

Cohere prices for the enterprise in three shapes: custom contracts for its North and Compass platforms, reserved capacity on Model Vault at $3.00 to $7.00 an hour or $2,000 to $4,300 a month per instance, and pay-as-you-go tokens behind a Production API key. A four-way toggle sorts the page by buying job, and the FAQ carries the billing mechanics, down to an invoice at month end or at $250 of outstanding balance.

Deepgram's page opens on "Free to start. Priced to scale." and three cards: Pay as you go with a Free $200 credit, Growth at $4,000+ / year in pre-paid credits for up to 20% off, and Enterprise at Custom pricing. A tabbed "Rates by product" section names the meter for each product, per second of audio, per 1,000 characters, or per minute of connection time, and a sixteen-question FAQ carries the billing mechanics, from hourly rates to the 10% overage premium.

ElevenLabs sells monthly credit bundles across seven plans, from Free with 10k credits per month to Business at $990 per month with 6M credits and 10 seats, split into an individual row and a For Businesses row. One credit currency meters speech, music, dubbing, image and video, and the page translates it into minutes of audio twice: in a calculator that recommends a plan and in a table that prints the price of an extra minute for every tier. Seats arrive at Scale, so the model is credits first with team access layered on for businesses.

Exa prices its search API per thousand requests, with a rate for each endpoint: Lite Search $4, Auto Search $7, Deep Search $12, Contents $1 per 1k pages, Monitors $2 and Answer $5, plus $1 per 1k for each result above ten and for AI page summaries. Agent runs offer metered compute at $0.10 per ACU and six effort levels: five fixed from $0.012 to $1.00 per request and an Ultra level billed on usage. The plan cards divide access, throughput and compliance, and the rate card below them carries the price.

Factory sells its software development agents as six numbered plans: three individual subscriptions at $20, $100 and $200 a month, a Teams plan at $60/mo per team plus $40/mo per seat, and Business and Enterprise shaped with sales. The individual ladder is one rate sold in three sizes: Plus is "~5x the usage of Pro" for five times the price and Max is "~10x the usage of Pro" for ten times, so every step holds at $20 per multiple of Pro. A five-question FAQ now opens on model support and names the billing edges a buyer checks, from using your own keys to exceeding limits.

Fireworks prices three product lines on three meters: serverless inference per token, training per 1M training tokens, and on-demand deployments per GPU second. Serverless usage now draws down a pre-paid credit balance with Auto Reload and a monthly spend limit, and the training side carries a fine-tuning matrix that doubles cleanly at each step plus an eight-model Serverless Training API that meters prefill, cached prefill, sample and train tokens separately.

Gamma prices per seat across four plans, $0, $9, $18 and $90 per seat per month on annual billing, and pairs every paid seat with 'Unlimited AI creations' under fair use. Credits run underneath: Free starts with '400 credits at signup', and the footnote states that 'Premium features (Agent, advanced models, & API) require credits'. The page is compact, cards then an eleven-question FAQ, and every plan is bought online.

The Gemini Developer API page is a docs rate sheet with three access tiers on top: Free, Paid and Enterprise set what a developer can call, and every model below is metered per 1M tokens across service-tier tabs that now run Standard, Batch, Flex, Priority and Contributor. It prints the full cost structure, from cache storage per hour to grounding per request, and dates its price changes in the cell, such as the new Gemini 4 Argon flagship at $2.00 through November 30, 2026 and $4.00 starting December 1, 2026.

Grammarly sells three plans, Free, Pro and Enterprise, and measures the step between them in AI prompts: 100 a month on Free, 2,000 per member per month on Pro and unlimited on Enterprise. The page is built around the Pro trial, with '$12 USD' struck through beside a 'Try for free' chip and the trial terms written into the card. Enterprise is 'available through Superhuman Go', which places the top plan inside the wider Superhuman platform.

HeyGen prices AI video on credits, a monthly allocation spent across every feature, and packages its three individual plans around which AI models each one unlocks. Free caps output at '3 videos a month, watermarked', Creator at '$29 / mo' carries '600 credits' and 'Avatar V, with limited usage', and Pro at '$49 / mo' wears the 'Most Popular' badge, opens a dropdown of credit tiers and adds a 'PRO ONLY FEATURES' panel with Seedance 2.5 and Cinematic Avatars.

Hugging Face publishes four price lists on one page: seats at $9, $20 and $50 per month, storage from $8 to $18 per TB with volume breaks, Spaces hardware by the hour from FREE, and Inference Endpoints from $0.033 per hour. PRO now includes '$2 of included inference credits monthly', a dollar-denominated allowance beside its storage multiples. The seat buys governance and collaboration, while the meters carry the cost of storing and running models, so the account scales with the AI workload it hosts.

Jasper sells two plans: Pro at $59 per seat per month on the yearly term, scoped to one seat with 2 Brand Voices, 5 Knowledge assets and 3 Audiences, and Business at custom pricing with unlimited customization and team features. The FAQ adds the usage layer: GEO Hub reports and Agents 'run on credits, which sit alongside seat-based pricing on the Business plan', a seat-plus-credits model for AI work.

Mistral's pricing page sells Vibe through plans that are bands of usage: Free is a capped entry, Pro at '$14.99 /mo' includes '$25.5 /mo in API credits', and Team at '$24.99 /user/mo' adds 'Even higher limits.' and workspace controls. Allowances written as multiples of Free, an included API credit allowance and a pay-as-you-go row tie each subscription to the token meter the FAQ sets out.

OpenAI's API tab is a rate card for four models, GPT-6 Astra, GPT-6.1 Sol, GPT-6.1 Terra and GPT-6 Luna, each priced per million input, cached input and output tokens, from $10.00 down to $0.10 per 1M input tokens. A 'Choose your processing mode' switch reprices every card in place, 'Batch -50%' or 'Data residency +10%', and output runs at five times input on every model.

OpenRouter prices its router as a platform fee on top of pass-through inference: N/A on Free, 5.5% on Standard, 8% on Business, and 'Fee discounts available' on Enterprise. The page opens straight into a four-column comparison table with the fee as its first row, because for a router the margin on other providers' models is the product. Enterprise now sets its BYOK allowance as 'Custom' and carries Private Endpoints as an 'Add on'.

Perplexity's Enterprise view prices per seat with a credit allowance attached to each one: Enterprise Pro at '$34 / month per seat when billed annually' with '500 monthly Computer credits/seat', and Enterprise Max at $271 with '15,000 monthly Computer credits/seat'. A third card for 'Additional security and governance' reads 'Custom pricing', and an organization can mix Pro and Max seats in one account.

Replicate publishes a pure rate sheet: public models bill per output or per second of hardware, private models bill for every second an instance is online, and a hardware table prices each SKU from cpu-small at '$0.09/hr' to 8x H200 at '$43.92/hr'. The opening two sentences explain both meters before any number appears.

Runway sells monthly credit allowances on four individual plans, Free, Standard at $12, Pro at $28 and Max at $76 a month billed annually, and translates every allowance into output on the card: '625 credits monthly = 52s of Gen-4.5, 104s of Gen-4 Turbo, or 78 Gen-4 Images (1080p).' One credit currency covers Runway's own models and partner models such as Kling, Seedance and Nano Banana Pro.

Suno sells music generation on credits, with three plans that pair a credit allowance with download counts and commercial rights. Free refills '50 credits per day', Pro at '$8 /month' carries '2,500 credits per month' and the 'MOST POPULAR' badge, and Premier at '$24 /month' carries '10,000 credits per month' and Suno Studio, with annual billing selected by default at 'SAVE 20%'.

Synthesia sells AI video on a credit pool that sits on top of a free monthly allowance, and every paid card converts its credits into minutes of video, minutes of dubbing, survey responses and Syren videos. Basic is $0 with 500 credits a month, Starter is $14/mo ($168 billed yearly) with 15,000 credits a year, Pro is $59/mo ($708 billed yearly) with 72,000 credits a year at a selectable credit level, and Enterprise moves to custom credit volumes. The cards lead straight into a 'Compare plans' table, an API pricing band and an Enterprise block.

Together AI prices eight products on one page, each on the unit that drives its cost: per-token serverless rates across 29 chat models, reserved throughput in PTUs at $0.05 per minute, GPU hours banded by commitment length, post-training GPU hours billed per minute, sandbox vCPU and RAM hours, filesystem storage per GiB and fine-tuning per million tokens with a per-job minimum. The page shows its working throughout, from cached input rates inline to the stated assumptions behind the PTU estimator and the rule for counting fine-tuning tokens.

Weights & Biases pairs seats with usage: Pro 'Starts at $60/month, billed monthly' with up to 10 Model seats, then storage at $0.03/GB above 100 GB and Weave data ingestion at $0.10/MB above 1.5 GB a month. The model splits the seat by workload, Model seats for training and fine-tuning and Weave seats for GenAI applications, and a Cloud-hosted and Privately-hosted toggle settles deployment before any price appears.

Writer sells two plans, Starter for a team of up to five and Enterprise for the whole organization, and presents them through capability: two cards, a long matrix in seven blocks and a FAQ that states the model. Per the FAQ, Starter is 'monthly or annual per-seat plans with fixed credit limits', and Enterprise adds unlimited free users, a Lite seat for new users, and solution packs and services as the account grows.

fal prices two products on one page with rate sheets in place of plans: GPU hours for serverless compute across six SKUs, and model APIs billed per output unit across video, image, audio and 3D. The "Units per $1" column turns seconds, tokens, images, megapixels, characters, generations and credits into one number a buyer compares directly.

xAI publishes its rates on the Models page of its developer docs, where each product card pairs a capability description with its rate rows. Three products carry six billing units, tokens, minutes, characters, hours of audio, images and seconds of video, each matched to the workload it meters, and the page treats price as one more spec of the model, printed in the same block as the context window.