All tasks / Comparisons
Claude Opus 5.5 vs Claude Sonnet 5.5
Compared only on public benchmark results, across the 8 tasks where both are listed: Claude Opus 5.5 is a tier higher in 4 of 8 shared tasks, they share a tier in 2.
Claude Opus 5.5
Anthropic's Opus-tier Claude model for long-running agentic coding and knowledge work.
- Vendor
- Anthropic
- Weights
- Closed: API and apps only
- Released
- 22 Sep 2026
- API price
- $4 in / $20 out per 1M tokens
- License
- proprietary
- Last verified
- 18 Oct 2026
Claude Sonnet 5.5
Anthropic's mid-tier Claude model, described as its best combination of speed and intelligence.
- Vendor
- Anthropic
- Weights
- Closed: API and apps only
- Released
- 28 Sep 2026
- API price
- $2 in / $10 out per 1M tokens
- License
- proprietary
- Last verified
- 18 Oct 2026
Task by task
| Task | Claude Opus 5.5 | Claude Sonnet 5.5 | Higher tier |
|---|---|---|---|
| Chat & assistants Models | S1,508 on LMArena Text Arena (overall, style control) · Cloud | A1,475 on LMArena Text Arena (overall, style control) · Cloud | Claude Opus 5.5 |
| Search & research Models | A1,211 on LMArena Search Arena (style control) · Cloud | –No public result yet · Cloud | Claude Opus 5.5 |
| Writing Models | S1,519 on LMArena Text Arena: Creative Writing (style control) · Cloud | A1,465 on LMArena Text Arena: Creative Writing (style control) · Cloud | Claude Opus 5.5 |
| Documents & OCR Models | –No public result yet · Cloud | –No public result yet · Cloud | — |
| Coding Models | S1,806 on LMArena WebDev / Code Arena (overall) · Cloud | S1,768 on LMArena WebDev / Code Arena (overall) · Cloud | Same tier |
| Image understanding Models | S1,293 on LMArena Vision Arena (overall, style control) · Cloud | A1,266 on LMArena Vision Arena (overall, style control) · Cloud | Claude Opus 5.5 |
| Browser & computer agents Models | –No public result yet · Cloud | –No public result yet · Cloud | — |
| Workflow automation Models | A69.5% on AutomationBench-AA (Artificial Analysis) · Cloud | A71.8% on AutomationBench-AA (Artificial Analysis) · Cloud | Same tier |
A tier spans a range of scores, and small gaps on a leaderboard are often within its margin of error, so only a different tier counts as ahead. How we rank.
Same benchmark, both results
| Benchmark | Claude Opus 5.5 | Claude Sonnet 5.5 | Snapshot |
|---|---|---|---|
| Artificial Analysis Intelligence Index Composite index over Artificial Analysis' own evaluation suite, higher is better | 58 | 56 | 18 Oct 2026 |
| AutomationBench-AA (Artificial Analysis) Score: average share of each task's objectives completed across 657 business-workflow tasks in 40 simulated SaaS apps via REST APIs; a task with any guardrail (business-rule) violation scores zero, higher is better | 69.5% | 71.8% | 19 Oct 2026 |
| LMArena Text Arena (overall, style control) Arena score (Bradley-Terry rating from pairwise human votes, style-controlled default view), higher is better | 1,508 | 1,475 | 15 Oct 2026 |
| LMArena Text Arena: Creative Writing (style control) Arena score (Bradley-Terry rating from pairwise human votes, creative-writing prompts only, style-controlled view), higher is better | 1,519 | 1,465 | 15 Oct 2026 |
| LMArena Vision Arena (overall, style control) Arena score (Bradley-Terry rating from pairwise human votes on image + text prompts, style-controlled view), higher is better | 1,293 | 1,266 | 9 Oct 2026 |
| LMArena WebDev / Code Arena (overall) Arena score (Bradley-Terry rating from pairwise human votes on web-app generation), higher is better | 1,806 | 1,768 | 18 Oct 2026 |
| ParseBench (models) Overall: mean of five rule-based dimension scores (tables, charts, content faithfulness, semantic formatting, visual grounding) on ~2,000 human-verified enterprise pages, higher is better | 79.8% | 70.2% | 9 Oct 2026 |