Rangliste

Mathematik Rankings

Von Wettbewerbs- bis Frontier-Mathematik: AIME, MATH, FrontierMath.

Gesamt767 Coding469 Mathematik90 Reasoning483 Agentic89 Wissen482 Multimodal452 Coding-Agenten68 Wissenschaft96 Mehrsprachig12 Schreiben95 Chat & Assistenten95 Research99 Tool-Nutzung45 Langkontext116 Tempo472 Preis-Leistung471 Open Weights183

90 Modelle

#	Modell	Podium Score	AIME 2025	FrontierMath	MATH	Geschwindigkeit
01	Claude Mythos Preview AnthropicReasoningVorschau	100.0	—	—	—	80 t/s
02	MiMo V2.5 Pro XiaomiOffen	100.0	—	—	86.2%	65 t/s
03	GPT-6 Astra OpenAIReasoning	97.6	—	97.6%	—	82 t/s
4	Kimi K3.1 Pro Moonshot AIOffenReasoning	95.0	—	95%	—	45 t/s
5	GPT-6.1 Sol OpenAIReasoning	95.0	—	95%	—	75 t/s
6	MiMo V2.6 Pro XiaomiOffen	95.0	—	—	—	72 t/s
7	Claude 5.5 Sonnet AnthropicReasoning	94.2	—	94.2%	—	94 t/s
8	Gemini 4 Argon GoogleReasoningVorschau	93.5	—	93.5%	—	88 t/s
9	Muse Spark 1.3 MuseReasoning	90.0	—	—	—	75 t/s
10	Gemini 3.1 Pro Preview GoogleReasoningVorschau	89.0	—	89%	—	105 t/s
11	Kimi K3.1 Highspeed Moonshot AIOffenReasoning	88.0	—	—	—	130 t/s
12	Mistral Large 4 MistralReasoning	87.0	—	87%	—	92 t/s
13	Claude 5.5 Haiku AnthropicReasoning	86.5	—	86.5%	—	145 t/s
14	MiMo v2.6 Flash XiaomiOffen	85.0	—	85%	—	118 t/s
15	Grok 4 Heavy xAIVorschau	84.6	100%	—	—	50 t/s
16	NVIDIA Nemotron 3.5 Super-VL NVIDIAOffen	84.0	—	84%	—	110 t/s
17	Simba 3.2 IndependentReasoning	84.0	—	—	—	62 t/s
18	GLM-4.5 Air Zhipu AIOffenReasoning	82.0	91%	—	—	96 t/s
19	Qwen3.7 Max Alibaba	81.6	—	—	—	203 t/s
20	Gemini 3.1 Flash Lite Preview GoogleVorschau	80.0	—	80%	—	210 t/s
21	Claude Opus 4.6 Anthropic	79.6	99.8%	—	—	50 t/s
22	Ling 3.1 Flash Ling	78.0	86%	—	—	175 t/s
23	Gemini 3.1 Pro Google	76.0	—	—	—	136 t/s
24	Gemini 3 Pro Google	72.9	100%	—	—	120 t/s
25	Kimi K3 Moonshot AIOffenReasoning	70.9	—	—	—	37 t/s
26	GPT-5.6 Sol OpenAI	70.7	—	89%	—	72 t/s
27	GLM-5.2 Z.aiOffen	69.8	—	—	—	164 t/s
28	Gemini 3 Flash Google	69.6	99.7%	—	—	120 t/s
29	Qwen3.8 Flash Next AlibabaOffen	68.0	—	—	—	168 t/s
30	DeepSeek V4.1 Flash DeepSeekOffenReasoning	68.0	—	—	—	214 t/s
31	DeepSeek V4 Pro DeepSeekOffen	66.5	—	—	—	66 t/s
32	Step-5 StepFunReasoning	65.0	—	—	—	110 t/s
33	GLM-5.3 Flash Z.aiOffen	64.0	—	—	—	182 t/s
34	Qwen3.7 Plus Alibaba	62.0	—	—	—	54 t/s
35	DeepSeek V4 Flash DeepSeekOffen	62.0	—	—	—	113 t/s
36	Muse Spark Meta	59.8	—	—	—	80 t/s
37	Qwen3.6 Plus Alibaba	59.2	—	—	—	55 t/s
38	Claude Opus 4.8 Anthropic	58.7	—	—	—	50 t/s
39	GPT-5.6 Terra OpenAI	57.3	—	84.9%	—	127 t/s
40	Claude Opus 4.7 Anthropic	57.0	—	—	—	80 t/s
41	GPT-5.2 OpenAI	53.0	100%	40.3%	—	90 t/s
42	Seed 2.0 Pro ByteDance	47.2	98.3%	—	—	40 t/s
43	Kimi K2.6 Moonshot AIOffen	46.9	—	—	—	40 t/s
44	Qwen3.5 397B-A17B AlibabaOffen	45.3	—	—	—	71 t/s
45	Agnes 3.0 Flash AgnesReasoning	42.0	—	—	—	235 t/s
46	GPT-5.6 Luna OpenAI	40.3	—	78.6%	—	175 t/s
47	DeepSeek V4 Flash Vision DeepSeek	39.0	—	—	—	120 t/s
48	G9v3-39A5B AI9StarsOffen	39.0	—	—	—	60 t/s
49	G9v3-3B AI9StarsOffen	39.0	—	—	—	60 t/s
50	GLM-5.3 Zhipu AI	39.0	—	—	—	90 t/s
51	Grok 4.6 (low) xAI	39.0	—	—	—	59 t/s
52	Grok 4.6 (medium) xAI	39.0	—	—	—	59 t/s
53	Grok 4.6 (xhigh) xAI	39.0	—	—	—	62 t/s
54	HyperNova 60B MultiverseOffen	39.0	—	—	—	367 t/s
55	LFM2.5-2.6B Liquid AIOffen	39.0	—	—	—	60 t/s
56	LFM2.5-8B-A1B Liquid AIOffen	39.0	—	—	—	338 t/s
57	LongCat 2.0 LongCatOffen	39.0	—	—	—	42 t/s
58	Magistral Small 1.2 MistralOffen	39.0	—	—	—	60 t/s
59	Qwen3.8 27B (xhigh) AlibabaOffen	39.0	—	—	—	53 t/s
60	Qwen3.8 27B (low) AlibabaOffen	39.0	—	—	—	64 t/s
61	Qwen3.8 27B (medium) AlibabaOffen	39.0	—	—	—	62 t/s
62	Qwen3.8 27B (Non-reasoning) AlibabaOffenReasoning	39.0	—	—	—	62 t/s
63	Hunyuan Hy3 Tencent	35.8	—	—	—	69 t/s
64	GPT-5.4 OpenAI	32.1	—	47.6%	—	60 t/s
65	Claude Sonnet 4.6 Anthropic	30.7	—	—	—	50 t/s
66	GPT-5.5 OpenAI	25.7	—	35.4%	—	60 t/s
67	GPT-5.1 OpenAI	13.0	94%	—	—	60 t/s
68	GPT-5 OpenAI	10.9	94.6%	—	84.7%	60 t/s
69	Qwen3.8 Max Alibaba	5.0	—	—	—	55 t/s
70	Eleven Multilingual v3 ElevenLabs	0.0	—	—	—	200 t/s
71	FLUX.2 Pro Black Forest Labs	0.0	—	—	—	30 t/s
72	GPT-Image 2 OpenAI	0.0	—	—	—	25 t/s
73	Grok Imagine Quality xAI	0.0	—	—	—	34 t/s
74	ElevenLabs Scribe v2 ElevenLabs	0.0	—	—	—	240 t/s
75	Cartesia Ink Whisper Cartesia	0.0	—	—	—	260 t/s
76	FLUX.2 Dev Black Forest LabsOffen	0.0	—	—	—	28 t/s
77	Voxtral 3 TTS Mistral	0.0	—	—	—	165 t/s
78	Qwen Audio 3.1 TTS+ AlibabaOffen	0.0	—	—	—	140 t/s
79	FLUX.2 Max Independent	0.0	—	—	—	35 t/s
80	LTX-2.3 Pro Lightricks	0.0	—	—	—	18 t/s
81	Step Audio 2 StepFun	0.0	—	—	—	150 t/s
82	Kling 3.0 Pro Kwai	0.0	—	—	—	25 t/s
83	FLUX.2 Klein 9B Black Forest LabsOffen	0.0	—	—	—	55 t/s
84	Qwen3 TTS Flash Alibaba	0.0	—	—	—	230 t/s
85	Wan 2.2 A14B AlibabaOffen	0.0	—	—	—	15 t/s
86	ElevenLabs v4 Turbo Independent	0.0	—	—	—	220 t/s
87	Deepgram Flux Independent	0.0	—	—	—	260 t/s
88	FLUX.2 Klein 4B Black Forest LabsOffen	0.0	—	—	—	90 t/s
89	LTX-2.3 Fast Lightricks	0.0	—	—	—	60 t/s
90	DeepSeek V3.2 DeepSeekOffen	0.0	93.1%	—	—	60 t/s

Der Podium Score ist eine gewichtete Mischung aus fünf Ranglisten. Siehe Methodik für Details.

Model Name

Provider

Podium Score0.0

Context128K

Price (1M out)$0/M

Speed0 t/s

Architecture & Highlights

Model overview and workload profile.

Quickstart cURL

curl https://api.openrouter.ai/api/v1/chat/completions ...