Side-by-side comparison · models
Claude Opus 5.5 vs Claude Sonnet 5 (Free)
What we have checked about each, line by line. Every value shows where it came from and when we last confirmed it.
Models
Claude Opus 5.5
Claude Opus 5.5 is Anthropic's flagship model for demanding reasoning, coding, and long-horizon agentic work, succeeding Claude Opus 5.
Models
Claude Sonnet 5 (Free)
Everyday Claude agent model for coding, planning, browsing, and general work.
Checked facts
Line by line
| Measure | Claude Opus 5.5 | Claude Sonnet 5 (Free) |
|---|---|---|
| Input price | $4.00 per million tokens | Not checked yet |
| Output price | $20.00 per million tokens | Not checked yet |
| Price on amazon-bedrock | $0.200 per million cached input tokens | Not checked yet |
| Price on amazon-bedrock | $4.00 per million input tokens | Not checked yet |
| Price on amazon-bedrock | $20.00 per million output tokens | Not checked yet |
| Price on anthropic | $0.200 per million cached input tokens | Not checked yet |
| Price on anthropic | $4.00 per million input tokens | Not checked yet |
| Price on anthropic | $20.00 per million output tokens | Not checked yet |
| Price on azure | $0.200 per million cached input tokens | Not checked yet |
| Price on azure | $4.00 per million input tokens | Not checked yet |
| Price on azure | $20.00 per million output tokens | Not checked yet |
| Price on claude-on-aws | $0.200 per million cached input tokens | Not checked yet |
| Price on claude-on-aws | $4.00 per million input tokens | Not checked yet |
| Price on claude-on-aws | $20.00 per million output tokens | Not checked yet |
| Price on google-vertex | $0.200 per million cached input tokens | Not checked yet |
| Price on google-vertex | $4.00 per million input tokens | Not checked yet |
| Price on google-vertex | $20.00 per million output tokens | Not checked yet |
| Latest version | anthropic/claude-opus-5.5-20260921 | Not checked yet |
| Time to first token (ms) | 683,930 | 1,350 |
| Time to first token (ms) on amazon-bedrock | 3,461 | Not checked yet |
| Time to first token (ms) on anthropic | 2,689 | Not checked yet |
| Time to first token (ms) on azure | 2,440 | Not checked yet |
| Time to first token (ms) on claude-on-aws | 3,320 | Not checked yet |
| Time to first token (ms) on google-vertex | 2,903 | Not checked yet |
| Output speed (tokens a second) | 94.3 | 64.1 |
| Output speed (tokens a second) on amazon-bedrock | 69 | Not checked yet |
| Output speed (tokens a second) on anthropic | 79 | Not checked yet |
| Output speed (tokens a second) on azure | 76 | Not checked yet |
| Output speed (tokens a second) on claude-on-aws | 80 | Not checked yet |
| Output speed (tokens a second) on google-vertex | 80 | Not checked yet |
| Tokens served on OpenRouter in a day | 1.1T (for 10 Nov 2026) | Not checked yet |
| Tool calls on OpenRouter in a day | 5.7M (for 10 Nov 2026) | Not checked yet |
| Artificial Analysis Intelligence Index | 59.1 | Not checked yet |
| LMArena text rating (Elo) | 1,513 | Not checked yet |
| LMArena web development rating (Elo) | 1,814 | Not checked yet |
| LMArena vision rating (Elo) | 1,324 | Not checked yet |
| Design Arena rating (Elo) | 1,461 | Not checked yet |
| Gpqa diamond | 90.6% (for 22 Sept 2026) | Not checked yet |
| Frontiermath tiers 1 3 v2 private | 91.2% (for 22 Sept 2026) | Not checked yet |
| Simpleqa verified | 72.2% (for 22 Sept 2026) | Not checked yet |
| Arc agi 2 | 93.3% (for 22 Sept 2026) | Not checked yet |
| Apex agents | 73.5% (for 22 Sept 2026) | Not checked yet |
| Arc agi | 98.5% (for 22 Sept 2026) | Not checked yet |
| Chess puzzles | 67% (for 22 Sept 2026) | Not checked yet |
| Dtbench | 98.9% (for 22 Sept 2026) | Not checked yet |
| Ebr bench | 71.4% (for 22 Sept 2026) | Not checked yet |
| Frontiercode | 54.6% (for 22 Sept 2026) | Not checked yet |
| Frontiermath tier 4 v2 private | 95% (for 22 Sept 2026) | Not checked yet |
| Frontierswe | 62.3% (for 22 Sept 2026) | Not checked yet |
| Furniture assembly | 83.3% (for 22 Sept 2026) | Not checked yet |
| Lmca | 68.2% (for 22 Sept 2026) | Not checked yet |
| Mirrorcode | 77.4% (for 22 Sept 2026) | Not checked yet |
| Mystery game puzzles | 71% (for 22 Sept 2026) | Not checked yet |
| Proofbench | 100% (for 22 Sept 2026) | Not checked yet |
| Remote labor index | 21.3% (for 22 Sept 2026) | Not checked yet |
| Spireclimb | 3.3% (for 22 Sept 2026) | Not checked yet |
| Weirdml v3 | 49.8% (for 22 Sept 2026) | Not checked yet |
| Otis mock aime 2024 2025 | 100% (saturated benchmark: the top models all score near the ceiling) (for 22 Sept 2026) | Not checked yet |
A line appears only when at least one record has a value we confirmed from a source. Values we have not confirmed are left out, never guessed. Records of different types share fewer lines.
Add a third
Compare with one more record
How this works
Share it, or build your own
This page works for any two or three records in the directory, of any type. The link stays the same whichever order you name them in, so you can share it as it is. Change the records in the address, for example /directories/vs/claude-code-vs-cursor, or start from our editorial comparisons.