What Claude Sonnet 5.5 actually is
Claude Sonnet 5.5 is the second model in Anthropic’s Claude 5.5 family, following Opus 5.5. Anthropic describes it as a faster, lower-cost complement to Opus 5.5 and positions it for well-scoped everyday work, bug fixing, coding, and polished documents, slides, spreadsheets, and design work .
The release is not simply a renamed Sonnet 5. Anthropic says Sonnet 5.5 produces output more than 30% faster, uses tools more efficiently in many agentic tasks, and improves coding and long-horizon work. The company also says that although the per-token rate is unchanged from Sonnet 5, the newer model can cost up to 30% less per completed task because it often reaches the result with fewer tokens.
For developers, the model ID is
claude-sonnet-5-5.
Anthropic lists a
1 million token context window,
a normal maximum output of
128,000 tokens,
adaptive thinking, and a reliable knowledge cutoff of June
2026.
Official details: Introducing Claude Sonnet 5.5 and Claude Sonnet 5.5 model overview .
Who can access Claude Sonnet 5.5?
Anthropic says Sonnet 5.5 is available across its platforms. On Claude’s current pricing page, Sonnet access is listed for Free, Pro, Max, Team, and Enterprise . That does not mean every plan gets the same amount of usage: the model may be available while the amount you can use it differs substantially by plan.
| Access route | What Anthropic currently publishes |
|---|---|
| Claude Free | Sonnet is available, with the smallest included usage allowance and five-hour resets. |
| Claude Pro | Sonnet is available with more included usage, Claude Code, and weekly limits in addition to the five-hour session limit. |
| Claude Max | Sonnet is available with either 5x or 20x the Pro per-session allowance, depending on the Max tier. |
| Team and Enterprise | Sonnet is listed as available. Limits and billing depend on seat type and organization plan. |
| Claude API |
Available as
claude-sonnet-5-5,
billed by token rather than by a consumer subscription
allowance.
|
| Cloud platforms | Anthropic lists Amazon Bedrock, Google Cloud, Microsoft Foundry, and Claude Platform on AWS among supported platforms. |
Access references: Claude plans and pricing and Sonnet 5.5 platform availability .
Claude subscription pricing is separate from Sonnet 5.5 API pricing
For people using Sonnet 5.5 inside Claude, the subscription price buys access to Claude features and an included usage allowance. It does not buy a fixed bundle of Sonnet 5.5 API tokens.
| Plan | Published price and usage position |
|---|---|
| Free | $0. Sonnet access is included, but capacity is limited and resets every five hours. |
| Pro | $20 per month in the US, or $200 billed annually, shown by Anthropic as an effective $17 per month. Pro provides at least 5x more usage per five-hour session than Free. |
| Max 5x | $100 per month. Provides 5x the Pro per-session usage allowance. |
| Max 20x | $200 per month. Provides 20x the Pro per-session usage allowance. |
| Team Standard | $20 per seat per month when billed annually, or $25 monthly. Anthropic describes Standard seats as having more usage than Pro. |
| Team Premium | $100 per seat per month when billed annually, or $125 monthly. Premium seats provide 5x more usage than Standard seats. |
| Enterprise | Anthropic currently lists $20 per seat per month, billed annually, plus usage charged according to model and task. |
Regional pricing, taxes, mobile-store pricing, and organizational contracts can differ, so the live billing page should control a purchase decision.
Current plan prices: Claude pricing , Pro plan details , and Max plan details .
Claude Sonnet 5.5 API pricing
The standard Claude API rates for Sonnet 5.5 are the same headline input and output prices as Sonnet 5. The important cost change is efficiency: Anthropic says Sonnet 5.5 usually needs fewer tokens to do the same work, so the cost per completed task can fall even though the base token prices did not.
| Usage type | Price per 1M tokens |
|---|---|
| Input | $2.00 |
| Output | $10.00 |
| 5-minute cache write | $2.50 |
| 1-hour cache write | $4.00 |
| Cache read | $0.20 |
| Batch API | 50% discount on standard input and output rates. |
Anthropic’s announcement compares Sonnet 5.5 with Opus 5.5, which is priced at $4 input and $20 output per million tokens. That makes Sonnet 5.5 the lower-cost 5.5-family choice when the work does not require the heavier Opus tier.
API prices: Sonnet 5.5 model specifications and Claude API pricing .
What the new Claude usage limits actually mean
The phrase “Sonnet 5.5 usage limit” can be misleading. Anthropic does not currently publish a single fixed number of Sonnet 5.5 messages that every subscriber gets. Instead, consumer and seat-based plans use plan-level usage pools. How quickly you consume them depends on conversation length, attached files, the model, the features you use, and the complexity of the work.
Anthropic increased five-hour usage limits for Pro, Max, Team, and seat-based Enterprise plans when it launched Opus 5.5 on 22 September 2026. Those higher plan allowances are relevant when using Sonnet 5.5 too, but they are not a special Sonnet-only quota.
| Limit | How it works |
|---|---|
| Five-hour session limit | Usage replenishes on a rolling five-hour session window. Pro gets at least 5x Free per session; Max tiers scale from the Pro allowance. |
| Weekly limit | Paid plans also have a weekly allowance across models. It resets at a fixed account-specific day and time each week. |
| Shared usage pool | Claude on web, desktop, mobile, and Claude Code draw from the same plan allowance rather than giving you independent pools for each surface. |
| Variable consumption | Long conversations, large files, tool use, and more demanding model behavior can consume the allowance faster, so message counts are not fixed. |
| Extra restrictions | Anthropic says it may apply additional weekly, monthly, model, or feature limits to manage capacity and fair access. |
Practical interpretation: if two people both have Pro, one can hit the limit sooner simply because their sessions are longer, contain more files, or use more expensive features. A “messages remaining” estimate from another user is therefore not a reliable Sonnet 5.5 limit for your account.
Usage rules: Claude pricing FAQ , usage-limit best practices , and Anthropic’s Opus 5.5 announcement describing the five-hour limit increase .
What happens when you hit the limit?
For a normal included-usage limit, the simplest option is to wait for capacity to reset. Paid users can also check Settings → Usage to see the current five-hour session bar, weekly usage, and the next weekly reset time.
-
Check which limit you reached. The five-hour session allowance and the weekly allowance are separate constraints.
-
Use a limit reset if one is available. Anthropic now gives eligible subscription users occasional resets that can restore either the five-hour or weekly allowance to full immediately.
-
Enable usage credits on supported paid plans if you need to continue. Once included plan usage is exhausted, credits let supported accounts keep working at standard API rates.
-
Move to a higher usage tier when the pattern is persistent. Max 5x and Max 20x are intended for people who repeatedly need more uninterrupted capacity than Pro.
A limit reset does not change the normal weekly reset schedule and cannot be undone after it is used. Anthropic says available resets and expiry times appear in Settings → Usage.
Official guidance: What is a limit reset? and Usage limit best practices .
API rate limits are a different system
If you use
claude-sonnet-5-5
through the Claude API, the consumer five-hour and weekly
subscription rules are not your main constraint. API
organizations instead have spend limits and throughput limits
such as requests per minute, input tokens per minute, and
output tokens per minute.
Anthropic’s current standard rate-limit table lists Sonnet 5.5 at up to 1,000 requests per minute, 2,000,000 input tokens per minute, and 400,000 output tokens per minute in the standard published tier table. Anthropic also notes that new organizations or accounts with limited usage history may begin below those standard limits, and an organization’s actual tier and limits should be checked in the Claude Console.
Sonnet 5.5 has its own rate-limit bucket rather than sharing the older combined Sonnet 4.x bucket. API rate limiting also replenishes continuously rather than resetting on the consumer plan’s five-hour schedule.
API limits: Claude Platform rate limits .
Why Claude may switch away from Sonnet 5.5 on some requests
Access to the model does not guarantee that every request will always be completed by Sonnet 5.5. Anthropic says Sonnet 5.5 is the first Sonnet model to launch with stronger cybersecurity fallbacks. A narrow set of higher-risk cybersecurity or frontier-model-development requests can fall back to Sonnet 5, while certain biology or distillation requests can be blocked directly.
Anthropic says routine software development and most life-sciences work are unaffected. If a conversation changes models, that behavior should therefore be treated separately from ordinary subscription usage exhaustion.
Fallback behavior: Why Claude switched models in your conversation with Sonnet 5.5 .