Claude or Gemini for writing in English?
Anthropic's best model against Google's best, on the same English prompts, in blind duels: the gap on the leaderboard, the direct record, and the kinds of writing where Gemini comes closest.
The verdict
Claude Opus 5.5 by Anthropic (thinking adaptive high) writes better English than Gemini 3.8 Flash by Google (thinking high) in our blind duels: rank 2 against rank 25 on the leaderboard, score 94.8 against 77.0.
Head to head, Claude takes 88.9% of the points over 18 direct duels (16 wins, 0 ties, 2 losses). Claude leads in 6 of the 10 kinds of writing where both have been judged, Gemini in 4.
| Model | Rank | Score | Duels |
|---|---|---|---|
Claude  Claude Opus 5.5 💡 | 2 | 94.8 | 1,522 |
Gemini  Gemini 3.8 Flash 💡 | 25 | 77.0 | 2,345 |
Each side is its best-ranked variant in this edition: Claude Opus 5.5 by Anthropic (thinking adaptive high) for Claude, Gemini 3.8 Flash by Google (thinking high) for Gemini. When a better variant of either arrives, this page follows it.
Kind of writing by kind of writing
Score and rank recomputed on the duels of each category alone; the leader is in colour. Click a kind of writing for its full ranking.
| Kind of writing | Claude | Gemini | Direct duels |
|---|---|---|---|
| Articles, blogs and editorial | 95.6 #2 | 85.2 #15 | 1–0–0 |
| Dialogue and screenplay | 91.0 #11 | 94.0 #3 | 0–0–1 |
| Explanation and popularization | 68.2 #29 | 83.9 #13 | — |
| Fiction and narrative | 94.2 #10 | 63.5 #38 | 1–0–0 |
| Marketing and sales | 43.4 #62 | 70.6 #26 | 0–0–1 |
| Personal communication | 98.6 #1 | 76.5 #28 | 11–0–0 |
| Professional and administrative | 61.1 #47 | 80.9 #10 | — |
| Social media and short formats | 96.7 #2 | 64.7 #35 | 1–0–0 |
| Speech and public speaking | 97.6 #2 | 80.0 #24 | 1–0–0 |
| Structured documents and reports | 93.0 #3 | 51.0 #56 | 1–0–0 |
Direct duels read wins–ties–losses for the first model. A dash means the model has not yet been judged in that category.
What each side stands for
“Claude” is the Anthropic model that ranks highest on the English leaderboard at export time, “Gemini” the Google cloud model that does. Google’s local Gemma models are a different family and have their own head-to-head with Qwen. The exact variants are named in the table above.
What the numbers measure
Blind duels only: two anonymous texts for the same English prompt, an AI judge picking the better one without seeing the names. The direct record counts the duels where these two variants met; the kind-of-writing table recomputes the ranking on one category at a time.
How to choose
On the duels judged so far, Gemini comes closest in explanation and popularization, and its weakest category is structured documents and reports. Personal letters and emails are where Claude pulls furthest ahead. Check the categories table for the kinds of writing you do most, then read the example pair below.
Writing example
Prompt
Write a concise status update (50–150 words) for a community garden committee, summarizing progress on the new toolshed project. Use a short structure with a heading and brief bullet-style points covering: the concrete foundation was poured last weekend, lumber has been ordered and arrives Friday, one volunteer carpenter is still needed for the roof, and the budget remains on track at $1,400 of the $1,600 allotted. End with the target completion date of August 20. Keep the tone factual and organized.
Published verdict: Claude (judged by GPT-6 Sol).
Claude · Claude Opus 5.5 💡
Toolshed Project: Status Update
- Foundation: The concrete foundation was poured last weekend and is curing as planned.
- Materials: Lumber has been ordered and is scheduled to arrive Friday.
- Volunteers: We still need one volunteer carpenter for the roof. Please contact the committee if you can help or know someone who can.
- Budget: Spending is on track at $1,400 of the $1,600 allotted, leaving $200 in reserve.
Target completion date: August 20
Gemini · Gemini 3.8 Flash 💡
Toolshed Construction: Progress Update
- Foundation: The concrete foundation was successfully poured last weekend and has cured properly.
- Materials: All required lumber has been ordered and is scheduled for delivery this Friday.
- Staffing: One additional volunteer carpenter is still needed to assist with the roof construction.
- Budget: Expenditures remain on track, with $1,400 utilized of the $1,600 total allotment.
Target Completion Date: August 20
