Claude Opus 5.5
Claude Opus 5.5 is served by 11 companies at prices ranging from $4.00 to $8.00 per million input tokens, a 2.0× difference for identical weights.
- cheapest host · Amazon Bedrock
- $4.00
- price spread across hosts
- 2.0×
- companies serving it
- 11
- intelligence index
- 58.4
Is it worth it?
Nothing in the catalogue is both cheaper and at least as capable. Claude Opus 5.5 is on the efficiency frontier at $8.86 per thousand support replies.
Every host, compared
The same weights, priced and configured differently by 11 companies. Precision matters: 4-bit builds measurably degrade output and cost the same.
| Host | $/1M in | $/1M out | Cached | Context | Precision | Uptime |
|---|---|---|---|---|---|---|
| Amazon Bedrockcheapest | $4.00 | $20.00 | $0.200 | 1.00M | unknown | 100% |
| Azure | $4.00 | $20.00 | $0.200 | 1.00M | unknown | 100% |
| $4.00 | $20.00 | $0.200 | 1.00M | unknown | 100% | |
| Claude Platform on AWS | $4.00 | $20.00 | $0.200 | 1.00M | unknown | 100% |
| Anthropic | $4.00 | $20.00 | $0.200 | 1.00M | unknown | 100% |
| Amazon Bedrock | $4.40 | $22.00 | $0.220 | 1.00M | unknown | 100% |
| Amazon Bedrock | $4.40 | $22.00 | $0.220 | 1.00M | unknown | 100% |
| Azure | $4.40 | $22.00 | $0.220 | 1.00M | unknown | n/a |
| $4.40 | $22.00 | $0.220 | 1.00M | unknown | 100% | |
| $4.40 | $22.00 | $0.220 | 1.00M | unknown | 100% | |
| Anthropic | $8.00 | $40.00 | $0.400 | 1.00M | unknown | 100% |
What it costs to run
- Marketing emails
- $9.46
- per 1,000 · short brief in, one email out
- Support replies
- $8.86
- per 1,000 · ticket plus help-centre context
- Lead enrichment
- $12.10
- per 1,000 · company page in, fields out
- Long-form drafts
- $43.12
- per 1,000 · outline in, full draft out
- Questions on docs
- $29.60
- per 1,000 · one large document, many questions
- Agent / tool loop
- $30.05
- per 1,000 · long context reused each step
Priced on the median host. Change the workload on the leaderboard to see how the ranking moves.
Specification
- Context, most hosts
- 1.00M · 1,500 pages
- Context range across hosts
- 1.00M to 1.00M
- Max output
- 128K
- Cache discount
- 96% off cached input
- Input types
- text, image, file
- Tool use
- supported
- Reasoning
- always on, billed as output
- Arena Elo
- 1383 across 8 categories
- Agentic Elo
- 1322
- AA Intelligence
- 58.4
Other models with the same host spread
GLM 5.3GLM 5.3 FlashDeepSeek V4.1 Flash 1017DeepSeek V4.1 FlashGLM 5.2Kimi K3DeepSeek V4 Flash 0731DeepSeek V4 Pro 0813
Host pricing and precision read live from OpenRouter’s endpoint API, compiled 2 November 2026. Free to quote with attribution. See all findings.