AnthropicVerified Active: September 18, 2026
Claude Sonnet 5API Pricing & Specifications
The flagship standard for software engineering, autonomous coding agents, and complex tool use.
📊 Core Specifications & Pricing Rates
Standard Input
$3.00 / 1M tokens
Standard Output
$15.00 / 1M tokens
Cached Input
$0.3000 / 1M tokens
Cache Write
$3.75 / 1M tokens
Context Window
1M (1,000,000 tokens)
Max Output Tokens
64,000 tokens
Supported Capabilities:✓ Vision✓ Function Calling / Tools✓ JSON Mode✓ Deep Reasoning
📦 Asynchronous Batch API Supported: Official provider documentation offers a 50% discount on non-streaming 24h batch requests.
Live Workload Cost Estimator: Claude Sonnet 5
Adjust token volumes to simulate your exact monthly API bill.
Estimated 60% repeated context
Estimated Total Monthly Cost
Compare All Models in Calculator→$201.75/ month
$0.00807 per API request(Saves $60.75/mo with caching)
💡 Estimated Monthly Workload Scenarios
AI Software Engineering Agent
$600.00/mo20,000 repository PR reviews & bug fixes (5,000 input, 1,000 output)
20,000 requests • 5000 in / 1000 out
Cached Enterprise Agent (75% cache hit)
$687.50/mo50,000 API queries (4,000 input, 600 output)
50,000 requests • 4000 in / 600 out
⚖️ Head-to-Head Competitor Comparisons
Claude Sonnet 5 vs. Gemini 3.8 Flash
Sonnet 5 leads on autonomous code refactoring and tool use, while Gemini 3.8 Flash provides 4x lower cost for high-throughput pipelines.
4x higher input cost than Gemini 3.8 FlashView Gemini 3.8 Flash →
Claude Sonnet 5 vs. GPT-5.6 Sol
Sonnet 5 is 25% cheaper on both input ($3 vs $4) and output ($15 vs $20) with top-tier agentic coding benchmarks.
1.33x cheaper than GPT-5.6 SolView GPT-5.6 Sol →
📜 Historical Price Changes & Updates
2026-08-15
Claude Sonnet 5 launched with 1M context window and enhanced computer use.
🔗 Verified Primary Sources
All pricing, token limits, and capabilities are cross-checked against official provider documentation and live API endpoints.