Claude Sonnet 5.5 Pricing, Benchmarks & Channels

Claude Sonnet 5.5 is a 2026-09-28 model from Anthropic featuring a 1M context window and extended reasoning capabilities. Lowest rate on Anthropic Official ($2 in / $10 out per 1M tokens).

⚖️ Compare in Radar Workbench 🧮 Monthly Cost Simulator
Input Price / 1M
$2
Output Price / 1M
$10
Cache Read / 1M
$0.1
Context Window
1M
Generation Speed
138 TPS
Access Grade
B · Global

F.A.R.E. Four-Dimension Score Breakdown

A · Accurate (Capability) 93/100 · Elo 1521 F · Fast (Speed & Latency) 138 Token/s · TTFT 44812ms R · Reliable (Stability & Access) Grade B · Global / Gateway E · Economical (Cost) $2 in · $10 out
Scenario Capability Breakdown (0–100):
💻 Coding
95/100
✍️ Writing
87/100
📚 Long-Doc RAG
88/100
👁️ Vision OCR
94/100
F.A.R.E. CAPABILITY RADAR
A · Accurate F · Fast E · Economical R · Reachable

Claude Sonnet 5.5 Multi-Platform Price Comparison (6 Verified Channels)

All prices are in USD per 1M tokens. Domestic developers can prioritize channels marked "Direct in CN" for compliant access and local billing.

Platform / Gateway Network Access Input ($/1M) Output ($/1M) Cache Hit ($/1M) Billing & Notes Link
Anthropic Official Lowest Price Global Network $2 $10 $0.1 Global cloud billing Visit Official →
302.AI Gateway Direct in CN $2 $10 $0.2 Direct CN access · RMB billing Visit Official →
OpenRouter Direct in CN $2 $10 $0.1 Direct CN access · RMB billing Visit Official →
Google Vertex AI Global Network $2 $10 $0.1 Global cloud billing Visit Official →
Google Vertex AI Global Network $2 $10 $0.1 Global cloud billing Visit Official →
AWS Bedrock Global Network $2.2 $11 $0.11 Global cloud billing Visit Official →

✅ Core Strengths & Ideal Workloads

  • Top-tier empirical capability score (93 index)
  • Fast generation speed (~138 tokens/sec)

⚠️ Limitations & Practical Caveats

  • Strict regional and payment controls on direct API accounts; use AWS Bedrock, Google Vertex AI, or OpenRouter for reliable global access.
  • Extended thinking generates internal reasoning tokens (~1.8x output multiplier)