Claude Opus 5.5 Launches on Amazon Bedrock, Anthropic's First Model in New 5.5 Family
Claude Opus 5.5, the first model in Anthropic's new Claude 5.5 family, is now live on Amazon Bedrock and Claude Platform on AWS. Anthropic claims the model does more with fewer tokens than Claude Opus 5, lowering average cost per task despite unchanged headline pricing tiers.
Anthropic's Claude Opus 5.5 is now available on Amazon Bedrock and Claude Platform on AWS, marking the first release in the company's Claude 5.5 model family. The model targets agentic coding, knowledge work, and long-running autonomous tasks.
What's new
According to Anthropic, Claude Opus 5.5 completes tasks using fewer tokens than its predecessor, Claude Opus 5. Combined with lower per-token pricing and cheaper cache reads, Anthropic claims this produces a lower average cost per task at scale. Exact per-token pricing figures were not disclosed in the announcement; teams should consult Amazon Bedrock's pricing page for current rates.
The model runs with adaptive thinking always enabled — Opus 5.5 determines its own reasoning depth per task rather than requiring developers to set manual "thinking budgets." Anthropic says developers can instead control behavior using an "effort" parameter. The model also surfaces intermediate progress — what it did, what it found, what it still needs — intended to make long-running agentic sessions easier to audit.
Claude Opus 5.5 is the first Opus-series model to ship with safety classifiers similar to those introduced in Claude Fable 5.1, covering biology, cybersecurity, and AI development risk categories. Anthropic states the model will refuse a higher rate of requests in these categories compared to prior Opus versions.
No context window size, parameter count, or third-party benchmark scores were disclosed in the announcement.
Availability and access
The model is accessible via Amazon Bedrock using the model ID global.anthropic.claude-opus-5-5, through US, EU, AU, JP, and Global Cross-Region Inference (CRIS) profiles on bedrock-runtime. It also runs on bedrock-mantle in US East (N. Virginia) and AP Southeast (Melbourne). Developers can invoke it through the Bedrock Invoke API, Converse API, or directly via the Anthropic Messages API using an AWS Bedrock authentication token. It is also available through Claude Platform on AWS in North America.
AWS notes that because Opus 5.5 is a reasoning model, API responses may include a "thinking" content block preceding the final text block — developers need to filter for the text block type rather than assume a fixed response index.
Use cases
Anthropic positions Opus 5.5 for software development teams running longer autonomous coding sessions, citing improved communication and explainability over Opus 5 as reasons the model is easier to review and trust. For knowledge work — drafting and revising long documents and reports — Anthropic claims the model requires fewer corrections than its predecessor. These are vendor claims; no independent benchmark data accompanied the announcement.
What this means
This release extends Anthropic's multi-cloud distribution strategy, putting a frontier-tier Opus model directly into AWS's Bedrock ecosystem alongside existing availability on other platforms. The efficiency claims — fewer tokens per task, lower cache-read costs — matter more to enterprise buyers than raw benchmark scores, since agentic workloads at scale are bounded by token cost, not just capability. The addition of mandatory safety classifiers with higher refusal rates signals Anthropic is willing to trade some raw compliance for risk mitigation in an Opus-tier model, a tradeoff enterprise customers running unattended, long-horizon agents will need to test against their own workflows before committing. The absence of published context window size, pricing, and benchmark numbers makes it difficult to compare Opus 5.5 directly against Opus 5 or competing frontier models until further documentation surfaces.
Related Articles
Google releases Gemini 4 Neon at $3/$15 per 1M tokens with a 1M-token context window
Google released Gemini 4 Neon, available today through the API at $3 per million input tokens and $15 per million output tokens. It is the fourth new Gemini 4 model in four weeks, with a context window of about one million tokens.
Anthropic Adds Experimental Code Mode to Claude Code v2.1.323, Chaining Tool Calls in JavaScript
Anthropic has added an experimental Code Mode to Claude Code v2.1.323. Claude can write JavaScript that chains Bash, Read and Edit calls in one step, and keep talking while long commands run.
Anthropic launches Linked billing beta, letting Claude API developers bill usage to a customer's own organization
Anthropic launched Linked billing in beta on November 17, 2026. Developers can bill Claude API usage in their applications to a customer's own Claude organization, choosing per request whether the customer or the developer pays.
Cloudflare releases open-weight Clairic 27B and new Clairic-flash 9B for data extraction
Cloudflare announced on November 18, 2026 an updated 27-billion-parameter Clairic and a new 9-billion-parameter Clairic-flash. Both have open weights on Hugging Face and are available as hosted Workers AI endpoints.
Comments
Loading...