Checked Oct 17, 2026
Anthropic
Claude Opus 5.5
Best for: Long-running agentic coding and knowledge work. Anthropic's suggested starting point.
- LMArena text #3 of 416
- Reasoning
- Reads text, images
- Input
- $4
- per 1M tokens
- Output
- $20
- per 1M tokens
- Cached input
- $0.20
- per 1M tokens
- Context window
- 1M
- tokens in one request
- Max output
- 125K
- tokens per reply
- Released
- Sep 22, 2026
- Knowledge cutoff
- Jun 2026
- Weights
- Closed
- API and apps only
Full 1M context at the standard price. Cached input = cache read price (0.05x base); 5-minute cache writes cost $5. Uses a newer tokenizer that produces roughly 30% more tokens for the same text than Claude 4.6 and earlier.
What would Claude Opus 5.5 cost you?
Change the numbers; the chart compares all 10 featured models at the same usage.
What it's good at
From Anthropic's own announcements and docs, linked.
- Anthropic's recommended model for most workloads, built for long-running agentic coding and knowledge work. source
- Anthropic calls it a major step up from Opus 5, and it costs less: $4 / $20 per million tokens against Opus 5's $5 / $25. source
- Reads values off dense charts, diagrams and screenshots much more precisely than earlier models, without extra tools. source
Watch out for
Limits and costs worth knowing before you build on it.
- Thinking is always on (default effort medium), and forced tool use returns an error.
- Safety classifiers can decline some cybersecurity and biology requests; Anthropic suggests a fallback model for those.
- Text between tool calls now arrives in thinking blocks, so apps that stream progress updates can go quiet until they're updated.
- Its tokenizer fits about 555,000 words in 1M tokens, against about 750,000 on Claude models before Opus 4.7.
Scores
Benchmark numbers are as published by Anthropic; we haven't rerun them. Makers use different setups, so don't compare one maker's numbers with another's. The leaderboard rank comes from public blind votes.
| Measure | Result | Source |
|---|---|---|
| LMArena text rank | #3 (1508) | Oct 15, 2026 |
| Terminal-Bench 4.0 (agentic coding, xhigh effort) | 66.4% | Anthropic |
| OSWorld 2.1 (computer use, partial credit, max effort) | 81.8% | Anthropic |
| Humanity's Last Exam (with tools, max effort) | 67.7% | Anthropic |
How to start
Try it in the app
The quickest way to see if it suits your work. Open Anthropic's app.
Get an API key
Sign up on Anthropic's developer platform and add billing; you pay per token as shown above. Read the docs.
Call the model
Use this model ID in your Anthropic API requests:
claude-opus-5-5
Host the app you build with it
Claude Opus 5.5 runs on Anthropic's servers, but your bot, agent or n8n workflow needs a home. A small VPS is usually enough, since the heavy work happens at Anthropic.
| Provider | From | Starting plan | Free tier | Our rating | Link |
|---|---|---|---|---|---|
| HostingerChecked | $2.99/mo | Premium (web hosting)20 GB SSD | No (30-day money-back guarantee) | Not rated | VisitOfficial site |
| DigitalOceanChecked | $4/mo | Basic Droplet 512 MiB1 vCPU · 0.5 GB RAM · 10 GiB SSD | Occasional promo credit for new users | Not rated | VisitOfficial site |
| VultrChecked | $2.50/mo | Cloud Compute Regular Performance, IPv6-only 0.5 GB1 vCPU · 0.5 GB RAM · 10 GB SSD | No free tier | Not rated | VisitOfficial site |
| Hetzner CloudChecked | $6.49/mo | CX23 (Cost-Optimized, Intel/AMD)2 vCPU · 4 GB RAM · 40 GB NVMe | No | Not rated | VisitOfficial site |
Compare with the other 9
| Model | Input / 1M | Output / 1M | Context | Arena |
|---|---|---|---|---|
| gpt-oss-120bOpenAI | $0.15 | $0.60 | 128K | #215 |
| Gemini 3.8 FlashGoogle | $0.75 | $3.75 | 1M | #9 |
| DeepSeek V4 ProDeepSeek | $1.32 | $3.96 | 1M | #56 |
| Grok 4.7xAI | $2 | $6 | 500K | #89 |
| Gemini 4 ArgonGoogle | $2 | $10 | 1M | #1 |
| GPT-6.1 SolOpenAI | $2 | $10 | 1.05M | #22 |
| Claude Sonnet 5.5Anthropic | $2 | $10 | 1M | #36 |
| Claude Opus 5.5Anthropic | $4 | $20 | 1M | #3 |
| Claude Fable 5.1Anthropic | $10 | $50 | 1M | #6 |
| GPT-6 AstraOpenAI | $10 | $50 | 1.05M | #30 |
Cheapest first, using 3 input tokens per output token.
Questions
How much does Claude Opus 5.5 cost?
Claude Opus 5.5 costs $4 per million input tokens and $20 per million output tokens at standard API rates, and $0.20 per million cached input tokens (US dollars, checked Oct 17, 2026). It is the 8th cheapest of the 10 models we feature.
How big is Claude Opus 5.5's context window?
1,000,000 tokens for the prompt, documents and conversation in one request. A single reply can be up to 128,000 tokens. How many words that holds depends on the model's tokenizer.
Is Claude Opus 5.5 open source?
No. Anthropic hasn't published its weights, so you use it through the Anthropic API or apps rather than on your own hardware.
When was Claude Opus 5.5 released?
Anthropic released it in Sep 22, 2026.
What is Claude Opus 5.5's knowledge cutoff?
Anthropic gives its training-data cutoff as Jun 2026. For later events it needs web search or documents you provide.
Sources
- Anthropic pricing, checked Oct 17, 2026
- platform.claude.com/docs/en/about-claude/models/overview
- platform.claude.com/docs/en/models/opus-5-5/overview
- platform.claude.com/docs/en/models/opus-5-5/whats-new-opus-5-5
- anthropic.com/claude-opus-5-5
Facts checked Oct 17, 2026. Model names belong to their makers; AskGPU isn't affiliated with Anthropic. Spotted something out of date? Tell us at hello@askgpu.com.