What each model API charges per million input and output tokens, with cached-input and batch prices, the rate-limit tiers (requests and tokens per minute, requests per day) and what unlocks each tier, free credits, and whether the balance is prepaid or billed, copied from OpenAI, Anthropic, Google, Mistral, xAI, DeepSeek, Groq, Together, Fireworks, OpenRouter, Cohere and Cerebras on the check date.
Free usage tier available to users in an allowed geography, with a $100/month usage limit. No trial credit amount is stated.
Build · $5Unlocked at $5 in total credit purchases
$10.00 per 1M tokensGPT-6 Astra (gpt-6-astra) input, Standard
$50.00 per 1M tokensGPT-6 Astra (gpt-6-astra) output, Standard
$1.00 per 1M tokensGPT-6 Astra (gpt-6-astra) cached input, Standard
Hard stopUsage is paid from prepaid credits bought in advance. If requests fail because the organization has used up its prepaid credits, add credits. Optional hard spend limits make affected requests return a 429 error.
Oct 30, 2026
Plans
Free
Custom
Qualification: user must be in an allowed geography. Usage limit: $100 / month."Free | User must be in an allowed geography | $100 / month"
A slow_down error can happen even within your RPM and TPM limits if traffic grows too quickly. Above 1M TPM, increase traffic by no more than 50% every 15 minutes.
Hard stopPrepaid credits with optional auto-reload. API requests stop when the credit balance runs out unless you buy credits or turn on auto-reload. Each tier also has a monthly spend cap: once it is reached, API usage pauses until 00:00 UTC on the first day of the next month unless a higher limit is granted. Organizations invoiced through Anthropic sales are billed for usage as usual.
Oct 30, 2026
Plans
Start tier
Custom
Monthly spend cap of $500 USD"| Start | $500 USD |"
Build tier
Custom
Monthly spend cap of $1,000 USD"| Build | $1,000 USD |"
Scale tier
Custom
Monthly spend cap of $200,000 USD"| Scale | $200,000 USD |"
Custom tier
Custom
No monthly spend cap; limits are arranged with the account team."Organizations on the Custom tier have no monthly spend cap; limits are arranged with their account team."
5-minute cache write 1.25x base input; 1-hour cache write 2x base input; cache read 0.1x base input (0.025x on Claude Fable 5.1 and Claude Mythos 5.1; 0.05x on Claude Opus 5.5 and Claude Sonnet 5.5)
A Claude Haiku 5.5 request whose prompt is over 100,000 tokens pays higher prices. Prompt length counts all input tokens, including cache reads and cache writes.
When tools are used, the API adds a tool-use system prompt (for example 286 tokens on Claude Opus 5.5, Sonnet 5.5 and Haiku 5.5 with tool choice auto or none), plus tool definitions and tool blocks billed as tokens.
Fast mode (Claude Opus 5.5 at $8 / $40 per MTok) cannot be combined with the Batch API; caching and data residency multipliers stack on top of fast mode pricing.
Short-interval enforcement and acceleration limits
A per-minute limit may be enforced over shorter intervals (e.g. 60 RPM as 1 request per second), and sharp usage increases can trigger acceleration-limit 429s.
Claude Max 5x ($100), Max 20x ($200) and Team (up to $500 pooled) include monthly Claude API credits that expire each billing cycle with no rollover and do not count toward moving to a higher tier.
Per 1M tokensPrices are listed per 1 million tokens in USD, separately for input, output (including thinking tokens) and context caching.
Free tier with limited access to certain models and free input and output tokens, up to each model's free-tier rate limits. Content on the free tier is used to improve Google's products. Gemini 4 Argon is not available on the Free tier.
Free · $0Limited access to certain models; free input & output tokens; Google AI Studio access; content used to improve Google's products. Qualification: active project or free trial.
$2.00 per 1M tokens through November 30, 2026Gemini 4 Argon – input (Standard)
$4.00 per 1M tokens starting December 1, 2026Gemini 4 Argon – input (Standard) from December 1, 2026
$10.00 per 1M tokens through November 30, 2026Gemini 4 Argon – output incl. thinking tokens (Standard)
Hard stopNew users default to Prepay: usage is deducted from a prepaid credit balance, and when it hits $0 all API keys on the billing account stop working (HTTP 402) until credits are added. Optional auto-reload with a monthly auto-charge limit. Postpay accounts are charged at month end or when the tier spend cap is reached; after the tier cap is reached, service is paused until the next billing cycle.
Oct 30, 2026
Plans
Free
$0
Limited access to certain models; free input & output tokens; Google AI Studio access; content used to improve Google's products. Qualification: active project or free trial."For developers and small projects getting started with the Gemini API."
Paid – Tier 1
Custom
Set up and link an active billing account. Billing tier cap $250. Higher rate limits, context caching, Batch API, access to most advanced models, content not used to improve products."Upgrading from the Free Tier to the Paid Tier means linking a billing account and prepaying to add a minimum of $5 (or equivalent in other currencies) of credit"
Paid – Tier 2
Custom
Paid $100 + 3 days from first successful payment. Billing tier cap $2,000."Paid $100 + 3 days from first successful payment"
Paid – Tier 3
Custom
Paid $1,000 + 30 days from first successful payment. Billing tier cap $20,000 - $100,000+."Paid $1,000 + 30 days from first successful payment"
Enterprise
Custom
All features in Paid, plus optional dedicated support channels, advanced security & compliance, provisioned throughput, volume-based discounts, ML ops and model garden, via Gemini Enterprise Agent Platform."For enterprise deployments, powered by Gemini Enterprise Agent Platform."
$0.075 per 1M tokens through December 31, 2026 ($0.15 from January 1, 2027); storage $0.50 / 1M tokens per hour through December 31, 2026 ($1.00 from January 1, 2027)
Output prices include thinking tokens. For Gemini 4 Argon, every token processed is billed, including thinking tokens and internal tool system instructions and function declarations when using grounding.
Content on the Free tier is used to improve Google's products; on the Paid tier it is not. The invite-only Contributor tier is also used to improve products.
Explicit context caching carries a storage fee per 1M tokens per hour (e.g. $4.50 for Gemini 3.1 Pro Preview and Gemini 2.5 Pro). Gemini 4 Argon supports implicit caching only.
Prepay billing is subject to overages during the approximately 10 minute billing pipeline latency; batch jobs and agents may consume credits beyond the balance.
After the cumulative billing-account total reaches the tier spend cap, service is paused for all linked projects until the 1st of the next month, unless an increase is requested.
Rate limits per project; RPD resets at midnight Pacific
Rate limits are applied per project, not per API key, and requests-per-day quotas reset at midnight Pacific time. Preview and experimental models have more restricted limits.
Per 1M tokensMost models are priced per million tokens, with input (prompts) and output (responses) counted separately. OCR is priced per 1,000 pages, speech models per minute, and tool APIs per call.
A free API tier with restrictive rate limits, intended for trying and exploring the API. The Free plan includes access to Mistral Studio. No free credit amount is stated.
API (Standard Tier, pay per use) · CustomBilled per million input and output tokens at per-model list prices
$0.7Mistral Large 4 input per 1M tokens
$0.07Mistral Large 4 cached input per 1M tokens
$2.1Mistral Large 4 output per 1M tokens
Not publishedNot found on the pages fetched. The billing docs page could not be fetched.
Oct 30, 2026
Plans
API (Standard Tier, pay per use)
Custom
Input and output tokens are counted separately and priced per model"Most models are priced per million tokens, with both input (your prompts) and output (responses) counted separately."
Batch
Custom
Asynchronous inference on large inputs at a 50% discount. Requests are queued and processed over a 24-hour period."Batching allows you to run asynchronous inference on large inputs in parallel, reducing compute costs while running large workloads at a 50% discount."
Priority Tier
Custom
Priority queueing, custom per-model rate limits, 99.5% uptime SLA. Requires account setup with Mistral."The multiplier is 1.75x standard list pricing, which is a 75% premium on input, output, and cached tokens."
In case you need to raise your usage limits, please feel free to contact us by utilizing the support button, providing details about your specific use case.not published
Per 1M tokensText API prices are per 1M tokens, in USD.
No free tier or trial credits are stated. Tier 0 is the default tier at $0 cumulative spend.not published
Pay-as-you-go (Build on your own) · CustomUsage-based; per-token rates in rates[]
$2.00 per 1M tokensgrok-4.7 input (prompt < 200k tokens)
$0.50 per 1M tokensgrok-4.7 cached input (prompt < 200k tokens)
$6.00 per 1M tokensgrok-4.7 output (prompt < 200k tokens)
Not publishedTier qualification counts prepaid credit purchases or paid invoices. Billing details are on a separate Manage Billing page.
Oct 30, 2026
Plans
Pay-as-you-go (Build on your own)
Custom
Access to all Grok models, usage-based pricing, automatically increasing rate limits"Access to all Grok models • Usage-based pricing • Automatically increasing rate limits"
Enterprise (Get extra support)
Custom
Dedicated onboarding support, custom rate limits, billing via monthly invoices, single sign-on and audit logging, data residency options"Dedicated onboarding support • Custom rate limits • Billing via monthly invoices • Single sign-on and audit logging • Data residency options"
Limits are set as requests per second (RPS) and tokens per minute (TPM); RPS is derived as RPM / 48 with a minimum of 2. No requests-per-day limit is stated.not published
Requests flagged as violating usage guidelines are still charged for generation. Violations caught before generation in the Responses API incur a $0.05 fee.
Per tokenTokens are the basic units used by models to represent natural language text, and also the units used for billing.
Not verifiednot published
Pay-as-you-go API · CustomPer-token prices not verified
2500deepseek-flash concurrency limit per account
500deepseek-v4-pro concurrency limit per account
No additional costCapacity expansion (higher concurrency)
Hard stopWhen the account balance runs out, requests return error 402 Insufficient Balance; funds are added on the Top up page.
Oct 30, 2026
Plans
Pay-as-you-go API
Custom
Access to deepseek-flash and deepseek-v4-pro; account balance is topped up on the Top up page."Cause: You have run out of balance. Solution: Please check your account's balance, and go to the Top up page to add funds."
deepseek-v4-flash and deepseek-v4-flash-vision-exp are still accepted but retired; requests are served by DeepSeek-V4.1-Flash-1017 and billed at the Flash price.
Search & Fetch billing in agent tools starts Oct 31, 2026
Search & Fetch calls via server-side tools in DeepSeek Harness, Claude Code, and Codex are free for DeepSeek API users until billing starts Oct 31, 2026, 16:00 UTC.
Per 1M tokensPrice per 1M tokens, listed separately for input and output; audio transcription priced per hour, text-to-speech per 1M characters.
Free tier with its own lower rate limits. Upgrading to the Developer tier requires a valid payment method. Groq occasionally offers promotional credits, such as at hackathons and special events.
Free · CustomNo price printed; Free tier
$0.15openai/gpt-oss-120b input per 1M tokens
$0.075openai/gpt-oss-120b cached input per 1M tokens
$0.60openai/gpt-oss-120b output per 1M tokens
Pay-as-you-goPostpaid: usage is invoiced monthly in arrears (with early invoices at progressive thresholds for new accounts) and charged automatically to the payment method on file. Optional monthly spend limits block API access when reached.
Oct 30, 2026
Plans
Free
Custom
Free plan rate limits (see rates)."Free Plan Limits"
Developer
Custom
Higher token limits, chat support, Flex service tier, Batch processing, Spend Limits."The Developer tier is designed for developers and companies who want increased capacity and more features with pay-as-you-go pricing."
Enterprise
Custom
Higher limits for select workloads and enterprise use cases; Enterprise-only models."Note that the limits shown below are the base limits for the Developer plan, and higher limits are available for select workloads and enterprise use cases."
Prompt caching limited to three models and not guaranteed
Prompt caching is only supported for openai/gpt-oss-20b, openai/gpt-oss-120b and openai/gpt-oss-safeguard-20b. Cache hits are not guaranteed, and the discount applies only on successful hits. Cached data expires after 2 hours without use.
Spend tracking has a 10-15 minute delay, so usage can exceed the limit by a small amount. When the limit is reached, API calls return 400 blocked_api_access. Spend limits are only available on paid plans.
Per 1M tokensChat, language, embedding, and rerank models are billed per input and output token; prices are listed per 1M tokens.
No free trial. Access requires a minimum $15 credit purchase.
Serverless (prepaid credits) · $15Minimum first credit purchase; $5 minimum for later purchases. Usage billed per token.
$3.00Kimi K3 input per 1M tokens
$0.30Kimi K3 cached input per 1M tokens
$15.00Kimi K3 output per 1M tokens
Hard stopFully prepaid. API access stops when the balance reaches zero and resumes after credits are added. Optional auto-recharge (card only) buys credits when the balance falls below a threshold and does not cap spend.
Oct 30, 2026
Plans
Serverless (prepaid credits)
$15
Prepaid credits usable for API requests, dedicated model inference, fine-tuning, evaluation jobs and all other Together AI services."The minimum is \$15 for your organization's first purchase and \$5 after that."
Scale / Enterprise contract
Custom
Billed under contract terms; purchased rate limits stay in place until the contract expires."If you have an active Enterprise or Scale contract, your purchased rate limits stay in place until the contract expires."
Batch is advertised at up to 50% off, but only listed models are discounted; others run at standard rates. The discount does not apply to dedicated model inference.
Per 1M tokensServerless inference is priced per token; prices are per 1 million tokens in US dollars, billed across input tokens, cached input tokens and output tokens.
$1 starter credit
Serverless — Standard · CustomPay per token; see rates
$3.00Ember-1 — Standard input per 1M tokens
$0.30Ember-1 — Standard cached input per 1M tokens
$15.00Ember-1 — Standard output per 1M tokens
Hard stopFireworks uses prepaid credits. When usage reaches 100% of the monthly spend limit, all API requests pause automatically; raise the limit to resume, and add credits if the balance is also depleted. Auto Reload can buy credits when the balance is low. Enterprise spend alerts do not pause service.
Oct 30, 2026
Plans
Serverless — Standard
Custom
Per-token pricing for input, cached input and output tokens."Serverless inference is priced per token."
Serverless — Priority
Custom
Priority tier is prioritized during high load to decrease the chance of being load shed."To **decrease the chance** of being load shed, you can use [Priority tier](https://docs.fireworks.ai/serverless/serverless-modes.md), which is prioritized durin"
Batch inference
Custom
Billed at 50% of serverless pricing on both input and output."**Batch inference** is billed at **50% of serverless pricing** on both input and output."
Unlimited / Enterprise
Custom
Enterprise accounts get higher upper bounds on adaptive rate limits automatically."A higher account [Spending Tier](https://docs.fireworks.ai/guides/quotas_usage/account-quotas.md#spending-tiers) correlates with higher upper bound rate limits;"
Auto Reload purchases credits when the balance is low; it does not change how much usage the account can accrue in the month. Adding credits does not raise or reset the monthly spend limit.
Limits are per account and per model, in tokens per minute, and start at a default that grows 25% at most every 15 minutes when usage is above 50% of the limit; they reset after more than 72 hours without traffic.
Per million tokens (prepaid USD credits)Each model and provider has its own price per million tokens, usually different for prompt (input) and completion (output) tokens; the request cost is deducted from prepaid USD credits.
Free plan: 25+ free models (:free variants) from 4 free providers, 50 requests/day (20 requests/minute); 1,000 requests/day once at least 10 credits have been purchased. New users receive a small free allowance (amount not stated).
Free · $0Platform fees: N/A
20 requests per minute, 50 requests per dayFree-model (:free) rate limit — less than 10 credits purchased (all time)
20 requests per minute, 1000 requests per dayFree-model (:free) rate limit — at least 10 credits purchased (all time)
$25,000 of list price inference per month with no fees, 5% fee afterBYOK fee-free allowance (Standard and Business)
Not published
Oct 30, 2026
Plans
Free
$0
25+ free models, 4 free providers, chat and API access, activity logs & export, 5 workspaces, 50 requests/day, community support. No prompt caching, budgets or auto-routing."| Rate limits | Free 50 requests/day |"
Standard
Custom
500+ models, 80+ providers, prompt caching, budgets & spend controls, auto-routing, management API key, admin controls, data-policy-based routing, 5 workspaces, BYOK $25,000/month list-price inference with no fees then 5%, rate limits passthrough from the provider(s), email support. Payment by credit card, crypto & more."Standard 5.5%"
Business
Custom
Everything in Standard plus advanced notifications, EU and US in-region routing, workload identity federation, 1,000 workspaces, signed DPA on request, email support."Business 8%"
Per 1M tokensGenerative models are priced per token, with input tokens (text sent to the model) and output tokens (text generated by the model) priced differently. Simple text may approach 1 token per word on average, while complex texts may use 3-4 tokens per word on average.
A free Trial API key is created automatically when you sign up. Trial calls are free but rate limited (20 requests/min for Chat models), limited to 1,000 API calls a month, and may not be used for production or commercial purposes.
Trial API key · $0Free
$2.5 / 1M tokensCommand A (command-a-03-2025) input per 1M tokens
$10 / 1M tokensCommand A (command-a-03-2025) output per 1M tokens
$0.0375 / 1M tokensCommand R7B (command-r7b-12-2024) input per 1M tokens
Pay-as-you-goProduction usage is billed pay-as-you-go. The bill is issued at the end of each calendar month, or when outstanding balances reach $250. A monthly spending limit can be set in the Billing & Usage section of the dashboard.
Oct 30, 2026
Plans
Trial API key
$0
Free API calls on all models and APIs. Rate limited and limited to 1,000 API calls a month. Not for production or commercial use."Trial keys (and prod keys on newer Chat model variants) are limited to 1,000 API calls a month."
Production API key
Custom
Charged pay-as-you-go and designed for production use at scale. You get it by completing the Go to Production workflow on the Billing and Usage page."API calls made from a Production API key will be charged on a pay-as-you-go basis. Production API keys are designed for production use at scale."
For both trial keys and production keys, Command A+ is free until rate limits are reached. It can be used in production through Cohere's Model Vault.not published
If you indicate your usage involves a sensitive use case, your production key may be rate limited the same as a trial key until Cohere's safety team approves it. Reviews take no longer than 72 business hours.
Chat responses report both total tokens and billed tokens. Tokens that Cohere adds under the hood, or special tokens the model is trained to output, are not charged.
Getting a Production key requires Owner privileges and completing the Go to Production workflow. Organizations do not get Production key access automatically.
Per 1M tokensPrice per million input tokens and per million output tokens, per model (Developer pricing).
Free Trial: one-time $5 credit after adding a verified payment method, expiring 30 days after it is granted, usable on all public models. No permanently free tier.
Free Trial · $0One-time $5 promotional credit; expires 30 days after activation
$0.35gpt-oss-120b input per 1M tokens
$0.75gpt-oss-120b output per 1M tokens
$0.99qwen-3.8-27b input per 1M tokens
Not publishedUsage is paid by purchasing credits (Pay as You Go); Developer users can use as many tokens as needed within their budget. Free Trial access stops when credits are used up or expire until credits are purchased.
Oct 30, 2026
Plans
Free Trial
$0
$5 in credits usable on all Shared Inference models, subject to Free Trial rate limits. For development, evaluation, and experimentation; not intended for production use."Add a valid payment method to receive a one-time $5 promotional credit. The credit expires 30 days after activation. You will not be charged or automatically en"
Developer (Pay as You Go)
Custom
Higher per-model rate limits; no hourly or daily token caps."Your first purchase moves you to the Developer tier with significantly higher rate limits and no hourly or daily token caps."
Enterprise
Custom
Rate limits provisioned per organization based on usage profile; custom pricing, dedicated support."Enterprise rate limits are provisioned per organization based on your usage profile."
Free: free to start. Limited access to certain models; free input & output tokens; Google AI Studio access; content used to improve Google's products. Qualification: active project or free trial. Checked 2026-10-30.
What is the Gemini API gemini 4 argon – input (standard)?
$2.00 per 1M tokens through November 30, 2026 (USD/1M tokens). Vendor wording: "$2.00 through November 30, 2026.". Checked 2026-10-30.
How much does Mistral AI API cost?
API (Standard Tier, pay per use): Billed per million input and output tokens at per-model list prices. Input and output tokens are counted separately and priced per model Checked 2026-10-30.
What is the Mistral AI API mistral large 4 input per 1m tokens?
Pay-as-you-go (Build on your own): Usage-based; per-token rates in rates[]. Access to all Grok models, usage-based pricing, automatically increasing rate limits Checked 2026-10-30.
What is the xAI Grok API grok-4.7 input (prompt < 200k tokens)?
Pay-as-you-go API: Per-token prices not verified. Access to deepseek-flash and deepseek-v4-pro; account balance is topped up on the Top up page. Checked 2026-10-30.
What is the DeepSeek API deepseek-flash concurrency limit per account?
Serverless (prepaid credits): $15 per usage. Prepaid credits usable for API requests, dedicated model inference, fine-tuning, evaluation jobs and all other Together AI services. Checked 2026-10-30.
What is the Together AI kimi k3 input per 1m tokens?
Each record is read from the vendor's pricing page, docs or help centre on the check date; every figure carries the vendor's exact sentence and a link to the page it came from. Where a vendor publishes no number the row says "not published" and quotes their wording instead. Numbers we saw only in third-party copies, forums or cached pages are kept out of the table. Reuse the JSON with a link back (CC BY 4.0).