All tasks / Comparisons
ChatGPT vs Claude
Compared only on public benchmark results, across the 10 tasks where both are listed: ChatGPT is a tier higher in 1 of 10 shared tasks, Claude in 3, they share a tier in 3.
ChatGPT
OpenAI's chat assistant on web, desktop and mobile, built on its GPT model family.
- Vendor
- OpenAI
- Kind
- Apps
- Runs
- Cloud
- Price
- Free tier · from $20/mo
- Skill level
- Beginner: works out of the box
- Open source
- No
- Last verified
- 18 Oct 2026
Claude
Anthropic's chat assistant on web, desktop and mobile, offering Claude Fable, Opus, Sonnet and Haiku models.
- Vendor
- Anthropic
- Kind
- Apps
- Runs
- Cloud
- Price
- Free tier · from $20/mo
- Skill level
- Beginner: works out of the box
- Open source
- No
- Last verified
- 18 Oct 2026
Task by task
| Task | ChatGPT | Claude | Higher tier |
|---|---|---|---|
| Chat & assistants Apps | Svia GPT-6.1 Sol · Cloud | Svia Claude Opus 5.5 · Cloud | Same tier |
| Search & research Apps | A48 on DeepResearch Bench (RACE, GPT-5.5 judge) · Cloud | Svia Claude Fable 5.1 · Cloud | Claude |
| Writing Apps | Avia GPT-6.1 Sol · Cloud | Svia Claude Opus 5.5 · Cloud | Claude |
| Translation Apps | –No public result yet · Cloud | –No public result yet · Cloud | — |
| Documents & OCR Apps | –No public result yet · Cloud | –No public result yet · Cloud | — |
| Data analysis Apps | C45.5% on SpreadsheetBench (V1 full, 912 tasks) · Cloud | C42.9% on SpreadsheetBench (V1 full, 912 tasks) · Cloud | Same tier |
| Image understanding Apps | Svia GPT-6.1 Sol · Cloud | Svia Claude Opus 5.5 · Cloud | Same tier |
| Voice agents Apps | Avia GPT-Live 1 · Cloud | –No public result yet · Cloud | ChatGPT |
| Browser & computer agents Agent apps & AI browsers | –No public result yet · Cloud | –No public result yet · Cloud | — |
| Slides & presentations Apps | Avia GPT-6 Astra · Cloud | Svia Claude Fable 5.1 · Cloud | Claude |
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 | ChatGPT | Claude | Snapshot |
|---|---|---|---|
| Mind2Web 2 (agentic search, private test set) Success rate: share of 130 long-horizon, time-varying agentic search tasks whose answer gets a perfect root score from the Agent-as-a-Judge rubric tree (leaderboard's default sort), higher is better | 28% | 10% | 19 Oct 2026 |
| SpreadsheetBench (V1 full, 912 tasks) Overall Pass@1 across 912 real-world spreadsheet questions and 2,729 test spreadsheets, higher is better | 45.5% | 42.9% | 18 Oct 2026 |