💬 Chat LLM-Rangliste

General conversation, instruction following, and helpfulness.

#	Modell	Gesamt	Arena ELO	IFEval	MMLU	Geschw.
1	Claude Opus 4 Anthropic	97.8	1385	94.2%	92.1%	45 t/s
2	Gemini 2.5 Pro Google	97.5	1398	92.3%	92%	85 t/s
3	Claude Sonnet 4 Anthropic	90.3	1340	90.1%	90.4%	79 t/s
4	GPT-4o OpenAI	89.2	1287	86.1%	88.7%	110 t/s
5	Grok 3 xAI	88.8	1330	87.8%	89.2%	55 t/s
6	Claude 3.5 Sonnet Anthropic	88.4	1272	86.5%	88.7%	82 t/s
7	Llama 4 Maverick Meta	87.5	1325	87.4%	89.5%	120 t/s
8	Gemini 2.5 Flash Google	86.7	1310	85.7%	86.5%	203 t/s
9	DeepSeek V3 DeepSeek	86.1	1318	86.7%	88.5%	68 t/s
10	Qwen 3 235B Alibaba	84.5	1305	85.2%	87.8%	95 t/s
11	Llama 3.1 405B Meta	82.1	1265	84.3%	87.3%	29 t/s
12	Gemini 2.0 Flash Google	80.5	1290	83.2%	85.1%	258 t/s
13	Llama 3.3 70B Meta	79.8	1258	83.5%	86%	65 t/s
14	Claude 3.5 Haiku Anthropic	79.2	1248	82.1%	84.1%	148 t/s
15	Mistral Large Mistral AI	76.5	1245	81.2%	84%	48 t/s
16	Qwen 2.5 72B Alibaba	76	1230	80.1%	85.2%	55 t/s
17	GPT-4o Mini OpenAI	75.8	1271	82.4%	82%	167 t/s
18	Grok 2 xAI	74.8	1240	78.5%	83.2%	60 t/s
19	Llama 3.1 70B Meta	74.3	1235	79.8%	83.7%	62 t/s
20	Mixtral 8x22B Mistral AI	65.2	1195	75.3%	77.8%	82 t/s
21	Command R+ Cohere	60.8	1185	72.5%	75.6%	52 t/s