whichdevtool?
Menu
StacksDirectory Pricing changes Glossary
VPS and cloud servers App hosting and PaaS Managed databases Transactional email APIs Email marketing and newsletters Domain registrars CDN, DNS and edge security Monitoring, uptime and error tracking Auth and identity Frontend and full-stack frameworks Backend frameworks UI component libraries AI model APIs File and object storage Payments and billing Background jobs, queues and workflows ORMs and database toolkits

AI model APIs

AI APIs let your app use models for text, images, audio and code.

Use for: Chatbots, document questions and generated content.

Amazon Bedrock Paid
Amazon Bedrock homepage

AWS service offering Claude, Amazon Nova, Llama, Mistral, DeepSeek, gpt-oss and other models on your AWS bill, with batch, flex, priority and provisioned modes.
Baseten Paid
Baseten homepage

Inference platform offering pre-optimized Model APIs for open models plus per-minute dedicated GPU deployments, with self-hosted and hybrid options.
Cerebras Paid
Cerebras Inference homepage

Inference API on Cerebras wafer-scale chips for a short list of open-weight models, prepaid by credit with no permanent free tier.
Claude API Paid
Claude API homepage

Anthropic's pay-per-token API for the Claude Fable, Opus, Sonnet and Haiku models, with batch and prompt-caching discounts and hard monthly spend caps per tier.
Cloudflare Freemium
Cloudflare Workers AI homepage

Serverless inference for open-weight models on Cloudflare's network, with 10,000 free Neurons a day and per-Neuron billing on the Workers Paid plan.
Cohere Freemium
Cohere API homepage

Enterprise-focused API for Cohere's Command, Embed and Rerank models, with a free but non-commercial trial key and dedicated Model Vault instances.
DeepInfra Paid
DeepInfra homepage

Low-price pay-as-you-go inference API for 100+ open-weight models, with a zero-retention policy, spending limits and GPU rental.
DeepSeek Paid
DeepSeek API homepage

DeepSeek's prepaid API for its V4.1 Flash and V4 Pro models, with very cheap cache hits and off-peak prices at half the peak rate.
Fireworks AI Paid
Fireworks AI homepage

Postpaid inference platform for open-weight models with Standard, Priority and Fast serving tiers, half-price batch jobs and monthly spend tiers.
Gemini API Freemium
Gemini API homepage

Google's developer API for Gemini models, with a free tier on most Flash models (prompts used to improve Google's products) and paid tiers with spend caps.
GroqCloud Freemium
GroqCloud homepage

Inference API on Groq's own LPU hardware for open-weight models such as gpt-oss and Qwen, with a free plan and published output speeds.
Hugging Face Freemium
Hugging Face Inference Providers homepage

One Hugging Face token routes requests to 200+ models on partner inference providers at the providers' own prices, with small monthly free credits.
Kimi API (Moonshot AI Open Platform) Paid
Kimi API (Moonshot AI Open Platform) homepage

Moonshot AI's international API for its Kimi K3 and K2 models, prepaid from a $1 minimum with rate limits that rise with cumulative recharges.
Microsoft Paid
Microsoft Foundry (Azure OpenAI) homepage

Azure's model platform: OpenAI GPT models as Azure OpenAI plus Claude and other models, billed to Azure with Global, Data Zone (EU/US) and Regional deployments.
Mistral API (Mistral Studio) Freemium
Mistral API (Mistral Studio) homepage

French model maker's API for Mistral Large, Medium, Small, Codestral, OCR and Voxtral models, most with open weights, and an EU inference option.
Novita AI Paid
Novita AI homepage

Low-cost API for 200+ open-weight text, image, audio and video models, plus GPU instances and agent sandboxes, with batch jobs at half price.
OpenAI Paid
OpenAI API homepage

Pay-per-token API for OpenAI's GPT-5.x, image, realtime voice and embedding models, with Batch, Flex and Priority processing modes priced separately.
OpenRouter Freemium
OpenRouter homepage

One API and one credit balance for models from OpenAI, Anthropic, Google and many open-model providers, with free models and provider fallback.
Replicate Paid
Replicate homepage

API for running community and official AI models, mostly image, video and audio, billed per second of hardware or per output.
SambaNova Freemium
SambaNova Cloud (SambaCloud) homepage

Inference API running open-weight models such as DeepSeek, gpt-oss, Gemma, Llama and MiniMax on SambaNova's own RDU chips, with a small free tier.
Scaleway Paid
Scaleway Generative APIs homepage

French cloud's OpenAI-compatible API for open-weight models hosted in Paris, priced in euros, with 1M free tokens for new customers.
Together AI Paid
Together AI homepage

Prepaid inference cloud for open-weight models such as DeepSeek, Qwen, Kimi, GLM, Llama and gpt-oss, plus dedicated endpoints, fine-tuning and GPU clusters.
Vertex AI (Gemini Enterprise Agent Platform) Paid
Vertex AI (Gemini Enterprise Agent Platform) homepage

Google Cloud's enterprise model platform, now branded Agent Platform, serving Gemini plus partner models such as Claude on a Google Cloud bill.
xAI Paid
xAI API (Grok) homepage

xAI's pay-per-token API for Grok models, with prepaid credits or monthly invoicing and a long-context surcharge above 200K tokens.

Tool Pricing ModelsFree tierExample price (per 1M tokens)Open-weight models
Amazon Bedrock aws.amazon.com · EU
Paid Claude, Amazon Nova, Llama, Mistral, DeepSeek, gpt-oss None Amazon Nova 2 Lite: $0.30 in / $2.50 out per 1M Yes: Llama, Mistral, DeepSeek, gpt-oss and others
Baseten baseten.co
Paid GLM-5.3, DeepSeek V4.1 Flash, Kimi K3, custom models No; free credits for new accounts DeepSeek V4.1 Flash: $0.30 in / $1.20 out per 1M Yes: GLM, DeepSeek, Kimi, plus any model you deploy
Paid gpt-oss-120b, GLM 4.7, Gemma 4 31B, Qwen 3.8 27B No; $5 trial credit for 30 days (card required) gpt-oss-120b: $0.35 in / $0.75 out per 1M Yes: gpt-oss, GLM, Gemma, Qwen
Claude API anthropic.com
Paid Claude Fable 5.1, Opus 5.5, Sonnet 5.5, Haiku 4.5 No free tier; small free credits for new users Claude Sonnet 5.5: $2 in / $10 out per 1M No (Anthropic's own models only)
Cloudflare Workers AI developers.cloudflare.com
Freemium Llama 3.x, Mistral 7B, DeepSeek V4 Pro, Granite; 50+ models 10,000 Neurons/day on Free and Paid plans Llama 3.1 70B: $0.293 in / $2.253 out per 1M Yes: Llama, Mistral, DeepSeek, Granite and others
Cohere API cohere.com
Freemium Command A family, Embed 4, Rerank 4, Aya Expanse Trial key: 1,000 calls/month, 20 req/min chat, non-commercial Aya Expanse 32B: $0.50 in / $1.50 out per 1M No (Cohere's own models)
DeepInfra deepinfra.com
Paid DeepSeek V4, Llama 3.3, Qwen3.6, Kimi K3, Nemotron 3, Mistral None; card or prepayment required DeepSeek-V4-Flash-0731: $0.06 in / $0.18 out per 1M Yes: DeepSeek, Llama, Qwen, Kimi, Mistral, Nemotron
DeepSeek API api-docs.deepseek.com
Paid DeepSeek-V4.1-Flash, DeepSeek-V4-Pro None; prepaid top-up balance deepseek-flash: $0.30 in / $1.20 out per 1M (peak) Serves DeepSeek's own models only
Fireworks AI fireworks.ai
Paid DeepSeek V4.1 Flash, GLM 5.3, Qwen 3.8, Nemotron 3.5 No; $1 starter credit, 10 RPM without a card DeepSeek V4.1 Flash: $0.30 in / $1.20 out per 1M Yes: DeepSeek, GLM, Qwen, Nemotron and others
Gemini API ai.google.dev
Freemium Gemini 3.8 Flash, 3.5 Flash-Lite, 3.1 Pro Preview, TTS, Live Free tokens on most Flash models; limits shown in AI Studio Gemini 3.8 Flash: $0.75 in / $3.75 out per 1M (to 2026-12-31) No
GroqCloud groq.com
Freemium gpt-oss 120B/20B, Qwen 3.6/3.8 27B, Whisper, Compound Per model, e.g. gpt-oss-120b 30 RPM, 1K RPD, 200K TPD GPT OSS 120B: $0.15 in / $0.60 out per 1M Yes: gpt-oss, Qwen, Whisper, Llama (enterprise)
Freemium 200+ models (DeepSeek, FLUX and more) via partners $0.10/month credits (free users), $2/month (PRO) Provider's own rate, no Hugging Face markup Yes: open models from the Hub via partner providers
Paid Kimi K3, Kimi K2.6, Kimi K2.7 Code None; $1 minimum recharge, $5 voucher at $5 recharged kimi-k3: $3 in / $15 out per 1M (cache hit $0.30) Serves Moonshot's own Kimi models
Microsoft Foundry (Azure OpenAI) azure.microsoft.com · EU
Paid GPT-5.4 family and older GPT models, Claude, other vendors None GPT-5.4 mini (Global): $0.75 in / $4.50 out per 1M Some partner models; OpenAI models are proprietary
Freemium Mistral Large 3, Medium 3.5, Small 4, Codestral, OCR, Voxtral $10/month API credits on the Free plan Mistral Medium 3.5: $1.50 in / $7.50 out per 1M Yes: Mistral Large 3, Medium 3.5, Small 4, Ministral
Novita AI novita.ai
Paid DeepSeek V4, Qwen3.x, GLM, Kimi, MiniMax, Llama, media None confirmed DeepSeek V4 Flash: $0.14 in / $0.28 out per 1M Yes: DeepSeek, Qwen, GLM, Kimi, MiniMax, Llama, Gemma
OpenAI API openai.com
Paid GPT-5.6, GPT-5.5, GPT-5.4 (mini, nano), gpt-image-2, realtime No free token allowance; moderation endpoint is free GPT-5.4 mini: $0.75 in / $4.50 out per 1M No (API serves OpenAI's own models)
OpenRouter openrouter.ai
Freemium OpenAI, Anthropic, Google, xAI, DeepSeek, Qwen, Llama and more Free models: 20 RPM, 50/day (1,000/day after $10) Provider list price + 5.5% fee on credit purchases Yes, alongside proprietary models
Replicate replicate.com
Paid FLUX, Nano Banana, gpt-image, Wan, DeepSeek R1, community Some models free to try, then billing required DeepSeek R1: $3.75 in / $10 out per 1M Yes: many community and open models, plus your own
Freemium DeepSeek V3.2, gpt-oss-120b, Gemma 4, Llama 3.3, MiniMax No card: 20 RPM, 20 requests/day, 200K tokens/day gpt-oss-120b: $0.22 in / $0.59 out per 1M Yes: DeepSeek, gpt-oss, Gemma, Llama, MiniMax
Scaleway Generative APIs scaleway.com · EU
Paid Mistral Medium 3.5, Qwen3.8, DeepSeek V4 Flash, Gemma 4, GLM 1M free tokens for new customers (one-off) deepseek-v4-flash-0731: €0.40 in / €0.80 out per 1M Yes: Mistral, Qwen, DeepSeek, Gemma, GLM, Whisper
Together AI together.ai
Paid DeepSeek V4, Qwen3.8, Kimi K3, GLM-5.3, Llama 3.3, gpt-oss None; $5 minimum prepaid purchase gpt-oss-120B: $0.15 in / $0.60 out per 1M Yes: DeepSeek, Qwen, Kimi, GLM, Llama, gpt-oss
Paid Gemini 3.x Flash, Flash-Lite, 3.1 Pro; Claude, partners None for model usage Gemini 3.8 Flash: $0.75 in / $3.75 out per 1M (to 2026-12-31) Partner and open models available alongside Gemini
Paid Grok 4.7, 4.5, 4.3, 4.20 (reasoning, multi-agent), grok-build None published on the pages read grok-4.3: $1.25 in / $2.50 out per 1M (<200K) No

Freemium means a free plan plus paid plans. Open a tool for its plans, prices and sources.

§1

How to choose an AI API

Terms: Tokens, Usage-based pricing and overage, Spend cap

  • Try the task on two or three models before choosing: quality differs by task more than benchmarks suggest.
  • Price by tokens in and out for your real prompts; output tokens usually cost several times more than input.
  • Set a spend limit or prepaid credit before you ship anything public. Usage-based bills grow with traffic and with loops in your own code.
  • A router such as OpenRouter lets you switch models with one API key, at a small markup; going direct can be cheaper and gives you the vendor’s own features first.

Questions

Which AI API has a free tier?

Several do, usually with low rate limits and sometimes with your prompts used for training. The "Free tier" column says what each offers as published.

§2

Pricing changes

  • Gemini 2.5 limited to existing users

    Gemini API · source ↗

  • DeepSeek V4 Pro kept after planned end date

    DeepSeek API · source ↗

  • DeepSeek-V4.1-Flash launched with lower prices

    DeepSeek API · source ↗

  • Claude Sonnet 5 keeps $2/$10 pricing

    Claude API · source ↗

  • DeepSeek adopts peak and off-peak pricing

    DeepSeek API · source ↗

All pricing changes →