AI preference rankings · 16 categories

Compare source rankings, uncertainty, provenance and dates for each category.

How to read this ranking

Arena aggregates anonymous comparison votes with a Bradley–Terry model. Scores measure user preference, not factual accuracy, safety or task success, and cannot be compared across categories. Score intervals and rank spreads come from the source. Overlapping ranges do not support a confident ordering.

Arena rank and interval methodology · Evaluation FAQ

🏆 Text overall →

Source cutoff 2026-09-13 · style_control

  1. claude-fable-5-high

    1,505.7 · Rank spread 1–7

  2. claude-opus-4-6-high

    1,504.6 · Rank spread 1–7

  3. claude-opus-4-7-high

    1,501.7 · Rank spread 1–10

Status: Source fetched

💻 Coding prompts →

Source cutoff 2026-09-13 · style_control

  1. claude-fable-5-high

    1,552.4 · Rank spread 1–10

  2. claude-opus-4-7-high

    1,551.8 · Rank spread 1–10

  3. claude-opus-4-6-high

    1,551.3 · Rank spread 1–10

Status: Source fetched

🌐 Web development →

Source cutoff 2026-09-23 · raw

  1. claude-opus-5.5-max

    1,818.4 · Rank spread 1–2

  2. gpt-6-astra-max

    1,792.2 · Rank spread 1–2

  3. claude-fable-5.1-max

    1,754.7 · Rank spread 3–3

Status: Source fetched

🧠 Hard prompts →

Source cutoff 2026-09-13 · style_control

  1. claude-opus-4-6-high

    1,533.1 · Rank spread 1–4

  2. claude-fable-5-high

    1,532.1 · Rank spread 1–6

  3. claude-opus-4-6

    1,526.7 · Rank spread 1–9

Status: Source fetched

📐 Math →

Source cutoff 2026-09-13 · style_control

  1. claude-fable-5-high

    1,526.3 · Rank spread 1–17

  2. claude-opus-5-high

    1,524.8 · Rank spread 1–17

  3. claude-opus-5-max

    1,523.1 · Rank spread 1–22

Status: Source fetched

✍️ Creative writing →

Source cutoff 2026-09-13 · style_control

  1. claude-fable-5-high

    1,504.1 · Rank spread 1–6

  2. claude-opus-4-6-high

    1,499.7 · Rank spread 1–6

  3. gemini-3.8-flash-high

    1,495.9 · Rank spread 1–17

Status: Source fetched

📋 Instruction following →

Source cutoff 2026-09-13 · style_control

  1. claude-opus-4-6-high

    1,513.5 · Rank spread 1–4

  2. claude-fable-5-high

    1,511.3 · Rank spread 1–7

  3. claude-opus-4-7-high

    1,502.7 · Rank spread 1–12

Status: Source fetched

💬 Multi-turn →

Source cutoff 2026-09-13 · style_control

  1. muse-spark-1.2 (xHigh)

    1,520.1 · Rank spread 1–28

  2. claude-fable-5-high

    1,517.6 · Rank spread 1–11

  3. claude-opus-4-6-high

    1,517.2 · Rank spread 1–10

Status: Source fetched

📜 Long queries →

Source cutoff 2026-09-13 · style_control

  1. claude-opus-4-6-high

    1,524.1 · Rank spread 1–6

  2. claude-fable-5-high

    1,521.4 · Rank spread 1–8

  3. claude-opus-4-6

    1,517.3 · Rank spread 1–10

Status: Source fetched

👁️ Vision →

Source cutoff 2026-09-13 · style_control

  1. claude-fable-5-high

    1,309.5 · Rank spread 1–8

  2. qwen3.8-max

    1,301.8 · Rank spread 1–14

  3. claude-opus-4-7-high

    1,300.9 · Rank spread 1–14

Status: Source fetched

📄 Documents →

Source cutoff 2026-09-13 · raw

  1. claude-opus-5-high

    1,516.3 · Rank spread 1–4

  2. claude-fable-5.1-max

    1,512.6 · Rank spread 1–7

  3. claude-opus-4-6-high

    1,507.3 · Rank spread 1–7

Status: Source fetched

🔎 Search →

Source cutoff 2026-08-24 · raw

  1. gpt-5.6-sol-xhigh

    1,257.3 · Rank spread 1–2

  2. claude-opus-4-6-search

    1,253.4 · Rank spread 1–2

  3. gpt-5.5-search

    1,242.1 · Rank spread 3–5

Status: Source fetched

🎨 Image generation →

Source cutoff 2026-09-24 · raw

  1. gpt-image-2.5-sunburst

    1,423.6 · Rank spread 1–1

  2. gpt-image-2.5-flare

    1,400.9 · Rank spread 2–2

  3. gpt-image-2 (medium)

    1,383 · Rank spread 3–3

Status: Source fetched

🖌️ Image editing →

Source cutoff 2026-09-21 · raw

  1. gpt-image-2.5-sunburst

    1,526 · Rank spread 1–1

  2. gpt-image-2.5-flare

    1,481.8 · Rank spread 2–2

  3. gpt-image-2 (medium)

    1,461.2 · Rank spread 3–3

Status: Source fetched

🎬 Video generation →

Source cutoff 2026-09-21 · raw

  1. gemini-omni-1.1-flash

    1,516.2 · Rank spread 1–4

  2. gemini-omni-flash

    1,512.9 · Rank spread 1–4

  3. flux-3-video

    1,493.1 · Rank spread 1–7

Status: Source fetched

🎞️ Image to video →

Source cutoff 2026-09-21 · raw

  1. minimax-h3

    1,495.4 · Rank spread 1–3

  2. gemini-omni-1.1-flash

    1,488.5 · Rank spread 1–5

  3. wan3.0

    1,479.8 · Rank spread 1–6

Status: Source fetched