Claude Sonnet 5 API Pricing
$2.00 input · $10.00 output per 1M tokens — the mid-tier workhorse between Haiku 4.5 and Opus 5.
Claude Sonnet 5 is Anthropic's mid-tier workhorse: $2.00 input and $10.00 output per million tokens, with cached input at $0.20. It costs 60% less than Opus 5 on output while covering the majority of production coding and agent workloads.
Price per 1M tokens
Claude Sonnet 5 · USD official list price
| Tier | Rate |
|---|---|
| Input | $2.00 |
| Output | $10.00 |
| Cache hit | $0.20 |
DeepSeek lists official prices in CNY; USD uses ≈ 6.9 exchange rate
Best for
$2.00 / $10.00
Mid-tier list price for input and output per million tokens.
$0.20 cached input
Cached input at one-tenth of the uncached rate.
60% cheaper than Opus
Sonnet 5's output costs 60% less than Opus 5's for the majority of workloads.
Production coding and agent loops that need more reasoning headroom than Haiku 4.5 but don't require Opus 5's ceiling — the default mid-tier default.
Billing notes
Published API pricing as of September 2026, standard tier, USD per million tokens. Anthropic offers a 50% Batch API discount for offline workloads.
Sonnet 5 vs the competition
Priced head-to-head with GPT-5.6 Terra ($2/$12), Sonnet 5 is slightly cheaper on output ($10 vs $12). DeepSeek V4 Flash remains far cheaper per token but without Anthropic's long-context product ecosystem.
Frequently Asked Questions
What is the price of Claude Sonnet 5 per million tokens?
Claude Sonnet 5 bills $2.00 per million input tokens and $10.00 per million output tokens, with cached input at $0.20 per million. It qualifies for Anthropic's 50% Batch API discount.
How does Sonnet 5 compare to Opus 5?
Sonnet 5 costs 60% less than Opus 5 on output ($10 vs $25) while covering the majority of production coding and agent workloads. It's the mid-tier workhorse between Haiku 4.5 and Opus 5.
Is Sonnet 5 good for agentic workflows?
Yes — it's the default mid-tier choice for agent loops that need more reasoning headroom than Haiku 4.5 but don't require Opus 5's ceiling. Model your token mix to confirm the trade-off.
Related model pricing
Compare this model against the rest of the lineup.
Estimate your exact cost
Model your own workload — input/output tokens, cache hit rate, and peak-hour share — against any model mix with the interactive pricing calculator.
Open the Pricing CalculatorRecommended Tools We ARE USING
We are using these tools ourselves for Development / Deployment. Check out for more details.
Opencode Go
You will receive $5 when subscribed, which directly offsets the 1st month's Go fee.
Could be the cheapest option for calling DeepSeek V4 Flash model in the world.
Vultr
RReceive $300 to test out VVultr platform
Referred user must be active 30+ days and spend $10–$25
Railway
Deploy anything without the complexity
Connect your repo, Railway handles the rest.
腾讯云 Tencent Cloud
Cloud servers, cloud databases, COS, CDN, SMS and other cloud products are on special offer now.
AWS high-quality alternative, we use COS as a replacement for S3
硅基流动 SiliconFlow
¥16 universal voucher for all platforms, valid for 180 days from the date of receipt
A comprehensive product matrix supporting the full-process implementation of AI applications.
Warp
One of the best AI terminal tools in the world
Interestingly, we can directly use various top-tier models in Warp without leaving the terminal.