Arene documents

De Wikiprompt, l’encyclopedie libre des prompts

Classement Elo communautaire, agrege chaque jour depuis les duels LMArena. L’arene generale d’abord, puis le detail par categorie.

Derniere mise a jour: 2026-10-13 · 44 modeles

Rang	Modele	Organisation	Score Elo	IC 95%	Votes	Licence
1	claude-opus-5-high	anthropic	1,516	1,509 / 1,523	8,794	Proprietary
2	claude-fable-5.1-max	anthropic	1,513	1,497 / 1,528	1,403	Proprietary
3	claude-opus-4-6-high	anthropic	1,507	1,501 / 1,514	27,929	Proprietary
4	claude-opus-4-6	anthropic	1,507	1,501 / 1,513	41,260	Proprietary
5	claude-fable-5	anthropic	1,496	1,488 / 1,504	8,497	Proprietary
6	claude-opus-4-7	anthropic	1,495	1,489 / 1,501	22,192	Proprietary
7	claude-opus-4-7-high	anthropic	1,495	1,488 / 1,501	21,957	Proprietary
8	gpt-5.5	openai	1,486	1,479 / 1,492	21,279	Proprietary
9	gpt-5.5-high	openai	1,484	1,478 / 1,491	20,944	Proprietary
10	gpt-5.6-sol-xhigh	openai	1,483	1,474 / 1,492	4,818	Proprietary
11	claude-sonnet-4-6	anthropic	1,482	1,476 / 1,488	57,813	Proprietary
12	claude-opus-4-8-high	anthropic	1,475	1,468 / 1,482	11,551	Proprietary
13	gpt-5.6-terra-xhigh	openai	1,472	1,463 / 1,480	5,787	Proprietary
14	gpt-5.4	openai	1,471	1,465 / 1,477	33,331	Proprietary
15	muse-spark-1.3-max	meta	1,471	1,452 / 1,489	1,006	Proprietary
16	gpt-6-astra-max	openai	1,468	1,453 / 1,482	1,621	Proprietary
17	claude-sonnet-5-high	anthropic	1,466	1,458 / 1,474	7,411	Proprietary
18	muse-spark-1.1	meta	1,465	1,456 / 1,474	4,885	Proprietary
19	claude-opus-4-8	anthropic	1,464	1,457 / 1,471	12,128	Proprietary
20	gemini-3.5-flash-medium	google	1,463	1,455 / 1,471	6,500	Proprietary
21	claude-opus-4-5-20251101	anthropic	1,462	1,452 / 1,473	7,981	Proprietary
22	gemini-3.5-flash-high	google	1,461	1,452 / 1,471	4,061	Proprietary
23	gpt-5.6-luna-xhigh	openai	1,457	1,449 / 1,465	5,766	Proprietary
24	gemini-3.6-flash-high	google	1,456	1,445 / 1,467	2,975	Proprietary
25	grok-4.6-high	xai	1,452	1,440 / 1,464	2,258	Proprietary
26	grok-4.5	xai	1,452	1,443 / 1,461	4,965	Proprietary
27	kimi-k2.6	moonshot	1,451	1,443 / 1,458	11,291	Modified MIT
28	claude-sonnet-4-5-20250929	anthropic	1,450	1,444 / 1,456	31,455	Proprietary
29	muse-spark	meta	1,444	1,426 / 1,462	1,084	Proprietary
30	qwen3.7-plus	alibaba	1,444	1,435 / 1,453	4,591	Proprietary
31	gemini-3.1-pro-preview	google	1,444	1,439 / 1,449	49,457	Proprietary
32	minimax-m3	minimax	1,435	1,427 / 1,443	6,314	MiniMax Community License
33	gemini-3-pro	google	1,434	1,425 / 1,443	10,769	Proprietary
34	kimi-k2.5-thinking	moonshot	1,430	1,423 / 1,437	21,569	Modified MIT
35	gemma-4-31b	google	1,425	1,417 / 1,432	12,326	Apache 2.0
36	gemini-2.5-pro	google	1,421	1,415 / 1,427	25,110	Proprietary
37	claude-haiku-4-5-20251001	anthropic	1,420	1,414 / 1,426	34,677	Proprietary
38	grok-4.20-beta-0309-reasoning	xai	1,416	1,409 / 1,423	20,948	Proprietary
39	glm-5v-turbo	zai	1,416	1,407 / 1,424	6,995	Proprietary
40	gemini-3-flash	google	1,413	1,404 / 1,422	7,158	Proprietary
41	gpt-5.2-high	openai	1,405	1,395 / 1,414	7,108	Proprietary
42	gpt-5.1	openai	1,403	1,394 / 1,413	8,244	Proprietary
43	gpt-5.5-instant	openai	1,403	1,395 / 1,411	8,497	Proprietary
44	gpt-5.2	openai	1,401	1,395 / 1,407	28,212	Proprietary

Voir aussi

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.