ai-api
Anthropic API pricing & features
API for Claude — frontier models for chat, tool use, agents, and long-context reasoning
REST API to Anthropic's Claude models. Supports streaming, prompt caching, extended thinking, tool use, vision, PDFs, batch processing, and citations. Distributed through Anthropic directly and via Amazon Bedrock, Google Vertex, and Azure AI.
Data sourced from vendor documentation · Last updated May 2026
Pricing
| Tier | Price | Notes |
|---|---|---|
| Opus 4.7 — usage | Custom | Input $5 / output $25 / cache-write $6.25 / cache-read $0.50 per 1M tokens. Batch = 50% off. |
| Sonnet 4.6 — usage | Custom | Input $3 / output $15 / cache-write $3.75 / cache-read $0.30 per 1M tokens. Batch = 50% off. |
| Haiku 4.5 — usage | Custom | Input $1 / output $5 / cache-write $1.25 / cache-read $0.10 per 1M tokens. Batch = 50% off. |
Limits
| Tier | Metric | Value | Notes |
|---|---|---|---|
| — | batch sla hours | 24 hours | Batch API completion SLA |
| — | cache ttl 1h | 1 hours | 1-hour cache TTL (beta) |
| — | cache ttl default | 5 minutes | Default prompt-caching TTL (ephemeral) |
| — | context window | 200000 tokens | Default context window across 4.x models |
| — | context window beta | 1000000 tokens | Opus 4.7 1M-token beta window (header gated) |
| — | max request batch mb | 256 MB | Max Batch API request size |
| — | max request files mb | 500 MB | Max Files API upload size |
| — | max request messages mb | 32 MB | Max Messages / Token Counting request size |
| haiku-4-5 | discount batch | 50 % off | Batch API discount vs real-time |
| opus-4-7 | discount batch | 50 % off | Batch API discount vs real-time |
| sonnet-4-6 | discount batch | 50 % off | Batch API discount vs real-time |
Features
Developer interfaces
| Slug | Name | Kind | Version |
|---|---|---|---|
| aws-bedrock | AWS Bedrock | rest | — |
| claude-code | Claude Code CLI | cli | — |
| console | Claude Console (Workbench) | ui | — |
| sdk-go | Go SDK | sdk | 0.x |
| gcp-vertex | Google Vertex AI | rest | — |
| sdk-java | Java SDK | sdk | 0.x |
| azure-ai | Microsoft Azure AI | rest | — |
| sdk-python | Python SDK | sdk | 0.x |
| rest | REST API (Messages + Agents) | rest | v1 |
| sdk-ruby | Ruby SDK | sdk | 0.x |
| sdk-typescript | TypeScript SDK (@anthropic-ai/sdk) | sdk | 0.x |
Related ai-api platforms
ai-api
AssemblyAI
Best-in-class speech-to-text API — Universal models, 99 languages, low-latency streaming
ai-api
Deepgram
Enterprise-grade speech-to-text + voice agents — Nova + Flux + Aura TTS
ai-api
ElevenLabs
Best-in-class AI text-to-speech + voice cloning + Conversational AI
ai-api
Google Gemini API
Gemini 2.5 Pro, Flash, Flash-Lite — multimodal + 2M context
ai-api
Groq
Fastest LLM inference — LPU-powered (300-1000+ tokens/sec)
Compare Anthropic API with
ai-api
Anthropic API vs AssemblyAI
Side-by-side breakdown.
ai-api
Anthropic API vs Deepgram
Side-by-side breakdown.
ai-api
Anthropic API vs ElevenLabs
Side-by-side breakdown.
ai-api
Anthropic API vs Google Gemini API
Side-by-side breakdown.
ai-api
Anthropic API vs Groq
Side-by-side breakdown.
ai-api
Anthropic API vs OpenAI API
Side-by-side breakdown.
ai-api
Anthropic API vs Replicate
Side-by-side breakdown.
ai-api
Anthropic API vs Together AI
Side-by-side breakdown.
ai-coding
Anthropic API vs Aider
Side-by-side breakdown.
ai-coding
Anthropic API vs Bolt.new
Side-by-side breakdown.
ai-coding
Anthropic API vs Claude Code
Side-by-side breakdown.
ai-coding
Anthropic API vs Cody
Side-by-side breakdown.
Staxly is an independent catalog of developer platforms. The link to Anthropic API above may be an affiliate link — Staxly may earn a commission if you sign up through it, at no extra cost to you. Pricing is verified at publication time — reconfirm on the vendor site before buying.