AI Model Profile
Claude Sonnet 5.5
Claude Sonnet 5.5 is Anthropic's speed-and-intelligence model in the Claude 5.5 family, released September 28, 2026, with a 1M-token context window and 128K maximum output.
- Family
- Claude
- Release date
- Unknown
- Status
- Current
- Context window
- 1M tokens
- Output limit
- 128K tokens
- API
- yes
- Open weights
- no
- Local/self-hosted
- no
- Pricing
- Pricing reviewed 2026-09-28. Claude API Standard: US$2 per million input tokens and US$10 per million output tokens, the same as Claude Sonnet 5. These are API usage rates, separate from Claude subscription pricing.
- Evidence state
- Source-verified
Benchmark Caveat
Benchmarks and provider capability notes are directional, not universal rankings. Results can shift with prompts, tool use, latency targets, pricing tier, safety filters, context length, and the workload mix a real team runs.
Anthropic reports 70.6% on Terminal-Bench 4.0, 81.3% on SWE-Bench Pro, 80.1% on OSWorld 2.1 (partial credit) and 1844 Elo on GDPval-AA v2.1, all at max effort. Artificial Analysis independently scored it 56 on its Intelligence Index v4.3.2 at max effort on September 28, 2026. Vendor figures are not a Kingy reproduction.
Best for
High-volume agentic coding, workflow automation, computer use, and document or spreadsheet work at Medium or High effort, where it lands within a few points of Claude Opus 5.5 at half the per-token price.
Skip if
Skip if the task needs max effort (Claude Opus 5.5 was cheaper per task there in independent testing), factual recall without search, or complex open-ended judgment, where Anthropic says Opus 5.5 remains clearly stronger.
Strengths
Anthropic lists Sonnet 5.5 with a 1M-token context window, 128K max output (300K on the Message Batches API with the output-300k-2026-03-24 beta header), adaptive thinking, fast comparative latency and a June 2026 reliable knowledge cutoff. Anthropic says it generates output 30%+ faster than Sonnet 5. Available on the Claude API, Amazon Bedrock, Claude Platform on AWS, Google Cloud and Microsoft Foundry, with zero data retention available.
Weaknesses
Token-heavy at max effort: Artificial Analysis measured $7.60 per Intelligence Index task, more than Claude Opus 5.5 ($5.98). Trails Opus 5.5 on factual recall and hard reasoning. thinking type "disabled" is rejected (the lowest setting is between_tools), and non-default temperature, top_p or top_k return an error.
Agent suitability
Useful for agent workflows: supports tool use, a 1M-token context window, structured outputs and computer use through the computer_toolset_20260801 toolset. Forced tool_choice (any or tool) is not supported.
Kingy AI take
Use this as a source-backed shortlist candidate, not a universal ranking. Re-check official provider docs and run a task-specific trial before production adoption.
Kingy AI Product Facts
Unknown
Current statusUnknown
- Company
- Not yet reviewed
- Primary job
- Not yet reviewed
- Audience
- Not yet reviewed
- Evidence
- Not yet verified
- Latest source check
- Unknown
See all tracked product factsPricing, platforms, dependencies, data claims, regions, and timeline
Sources and checks
- Coverage
- 0 of 88 tracked fields currently verified
- Sources
- 0
Source-check dates apply to the tracked facts. Editorial updates elsewhere on this page may have a different date.
Plans and pricing units
Not yet structured in this Kingy record.
Where it runs
- Availability
- Not yet reviewed
Stack and integrations
- Model / provider dependencies
- Not yet reviewed
- Integrations
- Not yet reviewed
Data and regions
- Vendor data-use claim
- Not yet reviewed
- Vendor retention claim
- Not yet reviewed
- Regions
- Not yet reviewed
Launch and latest material update
- Launch
- Not yet reviewed
- Latest material update
- Not yet reviewed
Coding notes
Strong on agentic coding at Medium or High effort. Effort levels were recalibrated from Sonnet 5, so re-run an effort sweep rather than carrying settings over.
Reasoning notes
Provider capability notes are useful but should be validated on representative prompts and tools.
Creative notes
Use a small creative test set before standardizing outputs for brand, media, or customer-facing work.
Research notes
Track release notes and model lifecycle notices because availability and aliases can change.
API pricing notes
Claude API pricing checked 2026-09-28; all amounts USD per million tokens. Batch: $1 input/$5 output. Standard prompt caching: $2.50 for 5-minute writes, $4 for 1-hour writes and $0.20 for cache hits/refreshes. Standard rates apply throughout the 1M-token context window. US-only inference adds 10% to all token categories. Uses the same tokenizer as Sonnet 5. Source review only; Kingy did not run paid requests or verify an invoice.
License notes
Commercial/API terms apply unless the linked official source states otherwise.
Hardware requirements
Cloud/API model; local hardware requirements are not published as a self-hosted path.
Official Model Links
Model Intelligence Research Map
Use these internal paths to move from this model profile into provider pages, static comparison pages, related Kingy records, and the broader AI launch graph. These are research paths, not rankings.