← Tous les instantanesVoir le classement en direct
Instantane des benchmarks: 2026-08-26
De Wikiprompt, l’encyclopedie libre des prompts
Photo figee des classements a cette date. Le classement en direct est sur /benchmarks.
← 2026-08-252026-08-27 →
Arene texte
| Rang | Modele | Organisation | Score Elo | Votes |
|---|
| 1 | claude-opus-5-high | anthropic | 1,504 | 27,610 |
| 3 | claude-opus-4-6-high | anthropic | 1,503 | 72,359 |
| 4 | claude-opus-4-6 | anthropic | 1,497 | 76,362 |
| 5 | claude-fable-5 | anthropic | 1,495 | 24,331 |
| 7 | claude-opus-4-7-high | anthropic | 1,490 | 60,203 |
| 9 | claude-opus-4-7 | anthropic | 1,483 | 61,304 |
| 10 | gemini-3.5-flash-high | google | 1,483 | 30,004 |
| 12 | gemini-3.1-pro-preview | google | 1,480 | 99,182 |
| 13 | gemini-3-pro | google | 1,479 | 41,491 |
| 14 | muse-spark-1.1 | meta | 1,478 | 20,242 |
| 18 | gemini-3.5-flash-medium | google | 1,475 | 28,343 |
| 21 | qwen3.5-max-preview | alibaba | 1,471 | 21,484 |
| 22 | gpt-5.5-high | openai | 1,471 | 59,545 |
| 23 | gpt-5.4-high | openai | 1,470 | 60,708 |
| 24 | ernie-5.1 | baidu | 1,468 | 37,137 |
Arene developpement web
| Rang | Modele | Organisation | Score Elo | Votes |
|---|
| 1 | claude-opus-5-max | anthropic | 1,691 | 8,116 |
| 2 | kimi-k3-max | moonshot | 1,674 | 4,546 |
| 3 | qwen3.8-max | alibaba | 1,669 | 3,212 |
| 4 | claude-opus-5-high | anthropic | 1,663 | 8,659 |
| 5 | grok-4.6-high | xai | 1,629 | 1,523 |
| 6 | claude-fable-5 | anthropic | 1,626 | 8,382 |
| 7 | gpt-5.6-sol-xhigh (codex-harness) | openai | 1,619 | 9,499 |
| 8 | glm-5.3-max | zai | 1,599 | 1,916 |
| 9 | qwen3.8-27b | alibaba | 1,595 | 2,464 |
| 10 | gemini-3.7-flash-high | google | 1,587 | 2,552 |
| 11 | glm-5.2-max | zai | 1,582 | 8,775 |
| 12 | deepseek-v4-pro-high-20260813 | deepseek | 1,582 | 2,869 |
| 13 | deepseek-v4-flash-high | deepseek | 1,579 | 3,537 |
| 14 | claude-opus-4-8-high | anthropic | 1,563 | 11,699 |
| 15 | claude-opus-4-7 | anthropic | 1,558 | 14,527 |
Arene vision
| Rang | Modele | Organisation | Score Elo | Votes |
|---|
| 1 | claude-fable-5 | anthropic | 1,328 | 9,562 |
| 2 | claude-opus-5-high | anthropic | 1,323 | 7,049 |
| 3 | claude-opus-4-7 | anthropic | 1,317 | 21,445 |
| 4 | claude-opus-4-7-high | anthropic | 1,316 | 21,122 |
| 5 | qwen3.8-max | alibaba | 1,315 | 6,743 |
| 6 | claude-opus-4-6-high | anthropic | 1,315 | 20,867 |
| 7 | gemini-3.5-flash-high | google | 1,313 | 7,586 |
| 8 | claude-opus-4-6 | anthropic | 1,311 | 25,169 |
| 9 | gemini-3.5-flash-medium | google | 1,308 | 7,533 |
| 10 | muse-spark | meta | 1,306 | 5,582 |
| 11 | gemini-3-pro | google | 1,305 | 13,134 |
| 12 | muse-spark-1.2 (xHigh) | meta | 1,304 | 1,853 |
| 13 | gemini-3.6-flash-high | google | 1,302 | 3,119 |
| 14 | gpt-5.4-high | openai | 1,297 | 22,515 |
| 15 | gpt-5.5 | openai | 1,295 | 20,727 |
Arene texte vers image
| Rang | Modele | Organisation | Score Elo | Votes |
|---|
| 1 | gpt-image-2 (medium) | openai | 1,380 | 69,194 |
| 2 | reve-2.1 | reve | 1,302 | 7,598 |
| 3 | muse-image | meta | 1,283 | 14,510 |
| 4 | reve-2.0 | reve | 1,270 | 14,650 |
| 5 | gemini-3.1-flash-image (nano-banana-2) [web-search] | google | 1,263 | 27,910 |
| 6 | qwen-image-3.0-pro | alibaba | 1,258 | 3,735 |
| 7 | seedream-5.0-pro | bytedance | 1,257 | 26,510 |
| 8 | mai-image-2.5 | microsoft-ai | 1,256 | 44,940 |
| 9 | gemini-3.1-flash-lite-image (nano-banana-2-lite) | google | 1,251 | 16,419 |
| 10 | gemini-3-pro-image-2k (nano-banana-pro) | google | 1,246 | 138,756 |
| 11 | gpt-image-1.5-high-fidelity | openai | 1,239 | 142,427 |
| 12 | gemini-3-pro-image-preview (nano-banana-pro) | google | 1,232 | 82,835 |
| 13 | grok-imagine-image-quality | xai | 1,228 | 48,786 |
| 14 | ideogram-4.0-quality | ideogram | 1,206 | 27,891 |
| 15 | qwen-image-2.0-pro-2026-06-22 | alibaba | 1,191 | 12,026 |
Arene edition d’image
| Rang | Modele | Organisation | Score Elo | Votes |
|---|
| 1 | gpt-image-2 (medium) | openai | 1,463 | 199,275 |
| 2 | grok-imagine-image-2.0 (low) | xai | 1,439 | 5,937 |
| 3 | mai-image-2.6-preview | microsoft-ai | 1,420 | 5,563 |
| 4 | muse-image | meta | 1,406 | 54,730 |
| 5 | mai-image-2.5 | microsoft-ai | 1,401 | 170,301 |
| 6 | seedream-5.0-pro | bytedance | 1,394 | 102,314 |
| 7 | grok-imagine-image-quality (20260519) | xai | 1,390 | 35,601 |
| 8 | gemini-3-pro-image-2k (nano-banana-pro) | google | 1,390 | 517,418 |
| 9 | chatgpt-image-latest-high-fidelity (20251216) | openai | 1,389 | 508,742 |
| 10 | gemini-3.1-flash-image (nano-banana-2) [web-search] | google | 1,386 | 117,943 |
| 11 | gemini-3-pro-image-preview (nano-banana-pro) | google | 1,385 | 519,140 |
| 12 | reve-2.1 | reve | 1,375 | 19,116 |
| 13 | gpt-image-1.5-high-fidelity | openai | 1,370 | 531,138 |
| 14 | reve-2.0 | reve | 1,358 | 17,508 |
| 15 | uni-1.1-max | luma-ai | 1,334 | 40,926 |
Arene texte vers video
| Rang | Modele | Organisation | Score Elo | Votes |
|---|
| 1 | gemini-omni-flash | google | 1,512 | 16,916 |
| 2 | flux-3-video | bfl | 1,494 | 1,291 |
| 3 | dreamina-seedance-2.0-720p | bytedance | 1,482 | 49,058 |
| 4 | dreamina-seedance-2.5-720p | bytedance | 1,477 | 1,337 |
| 5 | muse-video | meta | 1,457 | 2,176 |
| 6 | minimax-h3 | minimax | 1,453 | 3,081 |
| 7 | happyhorse-1.0 | aorizon | 1,428 | 22,113 |
| 8 | sora-2-pro | openai | 1,364 | 46,509 |
| 9 | veo-3.1-audio | google | 1,364 | 13,743 |
| 10 | veo-3.1-audio-1080p | google | 1,363 | 24,979 |
| 11 | veo-3.1-fast-audio | google | 1,361 | 39,441 |
| 12 | veo-3.1-fast-audio-1080p | google | 1,358 | 25,771 |
| 13 | veo-3-fast-audio | google | 1,347 | 25,219 |
| 14 | grok-imagine-video-720p | xai | 1,345 | 154,827 |
| 15 | wan2.7-t2v | wan | 1,344 | 17,769 |
Arene edition video
| Rang | Modele | Organisation | Score Elo | Votes |
|---|
| 1 | dreamina-seedance-2.5-720p | bytedance | 1,411 | 403 |
| 2 | minimax-h3 | minimax | 1,388 | 554 |
| 3 | dreamina-seedance-2.0-720p | bytedance | 1,359 | 3,634 |
| 4 | gemini-omni-flash | google | 1,358 | 1,868 |
| 5 | happyhorse-1.0 | aorizon | 1,307 | 3,115 |
| 6 | grok-imagine-video | xai | 1,262 | 13,281 |
| 7 | kling-o3-pro | kling | 1,257 | 7,125 |
| 8 | kling-o1-pro | kling | 1,200 | 10,104 |
| 9 | runway-gen4-aleph | runway | 1,183 | 9,817 |
Arene recherche
| Rang | Modele | Organisation | Score Elo | Votes |
|---|
| 1 | gpt-5.6-sol-xhigh | openai | 1,257 | 29,663 |
| 2 | claude-opus-4-6-search | anthropic | 1,253 | 134,699 |
| 3 | gpt-5.5-search | openai | 1,242 | 89,873 |
| 4 | claude-opus-4-7 | anthropic | 1,233 | 91,394 |
| 5 | claude-fable-5 | anthropic | 1,230 | 41,795 |
| 6 | ernie-5.1 | baidu | 1,227 | 3,788 |
| 7 | claude-sonnet-4-6-search | anthropic | 1,221 | 134,905 |
| 8 | grok-4.5 | xai | 1,213 | 31,505 |
| 9 | gemini-3.1-pro-grounding | google | 1,210 | 113,282 |
| 10 | gemini-3-pro-grounding | google | 1,207 | 37,024 |
| 11 | gpt-5.2-search | openai | 1,207 | 52,712 |
| 12 | claude-opus-4-8 | anthropic | 1,204 | 70,998 |
| 13 | grok-4.20-multi-agent-beta-0309 | xai | 1,204 | 109,553 |
| 14 | gpt-5.1-search | openai | 1,199 | 59,909 |
| 15 | gemini-3-flash-grounding | google | 1,198 | 149,334 |
Arene agents
| Rang | Modele | Organisation | Score Elo | Votes |
|---|
| 1 | Claude Opus 5 (High) | anthropic | | |
| 2 | Claude Opus 5 (Max) | anthropic | | |
| 3 | Claude Fable 5 (High) | anthropic | | |
| 4 | GPT 5.6 Sol (xHigh) | openai | | |
| 5 | Claude Opus 4.8 (High) | anthropic | | |
| 6 | Kimi K3 (Max) | moonshot | | |
| 7 | GPT 5.5 (xHigh) | openai | | |
| 8 | Claude Sonnet 5 (High) | anthropic | | |
| 9 | Claude Opus 4.7 (High) | anthropic | | |
| 10 | Claude Opus 4.7 | anthropic | | |
| 11 | DeepSeek V4 Pro (High) (0813) | deepseek | | |
| 12 | GPT 5.5 (High) | openai | | |
| 13 | Qwen3.8 Max | alibaba | | |
| 14 | Grok 4.5 | xai | | |
| 15 | GLM 5.2 (Max) | zai | | |
Arene documents
| Rang | Modele | Organisation | Score Elo | Votes |
|---|
| 1 | claude-opus-5-high | anthropic | 1,520 | 1,663 |
| 2 | claude-opus-4-6 | anthropic | 1,510 | 37,271 |
| 3 | claude-opus-4-6-thinking | anthropic | 1,506 | 24,973 |
| 4 | claude-fable-5 | anthropic | 1,504 | 5,792 |
| 5 | claude-opus-4-7 | anthropic | 1,498 | 18,990 |
| 6 | claude-opus-4-7-thinking | anthropic | 1,497 | 18,793 |
| 7 | gpt-5.5-high | openai | 1,485 | 16,816 |
| 8 | claude-sonnet-4-6 | anthropic | 1,483 | 54,427 |
| 9 | gpt-5.5 | openai | 1,480 | 17,286 |
| 10 | gpt-5.6-terra-xhigh | openai | 1,479 | 1,969 |
| 11 | gpt-5.6-sol-xhigh | openai | 1,479 | 1,152 |
| 12 | claude-opus-4-8-thinking | anthropic | 1,475 | 8,388 |
| 13 | muse-spark-1.1 | meta | 1,472 | 2,054 |
| 14 | claude-sonnet-5-high | anthropic | 1,470 | 3,601 |
| 15 | gpt-5.4 | openai | 1,470 | 29,809 |
Autres classements
LiveBench
Metrique: average score (0-100) · livebench.ai
| Rang | Modele | Organisation | Score | Votes |
|---|
| 1 | claude-fable-5-max-effort | | 83.4 | |
| 2 | gpt-5.6-sol-max | | 81.7 | |
| 3 | gpt-5.5-xhigh | | 80.8 | |
| 4 | claude-opus-5-max-effort | | 80.5 | |
| 5 | gemini-3.7-flash-high | | 79.9 | |
| 6 | smaug-agentic | | 79.7 | |
| 7 | qwen3.8-max | | 79.5 | |
| 8 | kimi-k3 | | 79.5 | |
| 9 | grok-4.6 | | 79 | |
| 10 | muse-spark-1.2-xhigh | | 78.9 | |
| 11 | gpt-5.4-xhigh | | 78.8 | |
| 12 | gpt-5.6-terra-max | | 78.6 | |
| 13 | deepseek-v4-pro-0813 | | 78.2 | |
| 14 | gemini-3.1-pro-preview-high | | 78 | |
| 15 | deepseek-v4-flash-vision-exp | | 77.7 | |
Aider polyglot (coding)
Metrique: pass rate 2 (%) · aider.chat/docs/leaderboards/
| Rang | Modele | Organisation | Score | Votes |
|---|
| 1 | gpt-5 (high) | | 88 | |
| 2 | gpt-5 (medium) | | 86.7 | |
| 3 | o3-pro (high) | | 84.9 | |
| 4 | gemini-2.5-pro-preview-06-05 (32k think) | | 83.1 | |
| 5 | o3 (high) | | 81.3 | |
| 6 | gpt-5 (low) | | 81.3 | |
| 7 | grok-4 (high) | | 79.6 | |
| 8 | gemini-2.5-pro-preview-06-05 (default think) | | 79.1 | |
| 9 | o3 (high) + gpt-4.1 | | 78.2 | |
| 10 | Gemini 2.5 Pro Preview 05-06 | | 76.9 | |
| 11 | o3 | | 76.9 | |
| 12 | DeepSeek-V3.2-Exp (Reasoner) | | 74.2 | |
| 13 | Gemini 2.5 Pro Preview 03-25 | | 72.9 | |
| 14 | o4-mini (high) | | 72 | |
| 15 | claude-opus-4-20250514 (32k thinking) | | 72 | |
Sources et methodologie
Les classements sont agreges depuis des evaluations communautaires ouvertes. Les scores Elo, intervalles de confiance et votes proviennent de LMArena (duels de modeles votes par la communaute, publies en dataset ouvert). Les prix et metadonnees de contexte proviennent de l’API publique OpenRouter. Wikiprompt prend un instantane quotidien et preserve l’historique ; nous ne menons aucune evaluation proprietaire.