IQ of SI
Deutsch ▾

Automatisierte Seite. Automatisch von KI verfasst und übersetzt und per Software mit den Quellen abgeglichen; keine menschliche Redaktion. SI-IQ ist ein Index auf Basis öffentlicher Benchmark-Ergebnisse, kein IQ-Test. So funktioniert es

GPQA Diamond (OpenRouter run): Ergebnisse von KI-Modellen

GPQA Diamond (OpenRouter run) ist ein Benchmark im Bereich Wissen, durchgeführt von OpenRouter. Für 117 Modelle in unseren Daten liegt ein veröffentlichtes Ergebnis vor.

Er ist Teil des SI-IQ-Index (Ladung k = 0,68, Zufallsniveau 25 %).

Evaluator
OpenRouter
Startseite des Benchmarks
openrouter.ai/benchmarks/gpqa-diamond

Ergebnisse

Quelle
1 Gemini 4 Argon Google DeepMind96,5 %140137–143 8 09.2026OpenRouter benchmark datasets
2 Gemini 3.1 Pro Google DeepMind94,8 %127124–130 9 02.2026OpenRouter benchmark datasets
3 Gemini 3.8 Flash Google DeepMind94,8 %130127–133 8 09.2026OpenRouter benchmark datasets
4 GPT-6 Astra OpenAI94,7 %150146–153 8 09.2026OpenRouter benchmark datasets
5 GPT-6.1 Sol OpenAI94,4 %148144–152 8 09.2026OpenRouter benchmark datasets
6 Gemini 3.7 Flash Google DeepMind93,9 %131128–134 8 08.2026OpenRouter benchmark datasets
7 GPT-5.5 OpenAI93,6 %138135–141 9 04.2026OpenRouter benchmark datasets
8 Gemini 3.6 Flash Google DeepMind93,0 %125122–128 8 07.2026OpenRouter benchmark datasets
9 Claude Sonnet 5.5 Anthropic92,9 %140137–143 8 09.2026OpenRouter benchmark datasets
10 Grok 4.6 xAI92,5 %126123–129 8 08.2026OpenRouter benchmark datasets
11 Gemini 3.5 Flash Google DeepMind92,4 %126123–129 9 05.2026OpenRouter benchmark datasets
12 GPT-5.6 Sol OpenAI92,3 %140137–143 8 07.2026OpenRouter benchmark datasets
13 GPT-6 Sol OpenAI92,1 %140137–144 8 09.2026OpenRouter benchmark datasets
14 Kimi K3offen Moonshot91,6 %127124–130 8 07.2026OpenRouter benchmark datasets
15 MiniMax-M3offen MiniMax91,5 %108102–114 5 06.2026OpenRouter benchmark datasets
16 GLM-5.3-Flashoffen Z.ai (Zhipu AI)90,9 %120117–124 7 08.2026OpenRouter benchmark datasets
17 Qwen3.7-Max Alibaba90,9 %125122–128 9 05.2026OpenRouter benchmark datasets
18 Muse Spark Openoffen Meta AI90,6 %127124–130 6 10.2026OpenRouter benchmark datasets
19 Claude Opus 5.5 Anthropic90,2 %144141–148 8 09.2026OpenRouter benchmark datasets
20 DeepSeek V4 Flash 0731offen DeepSeek90,0 %124121–127 8 07.2026OpenRouter benchmark datasets
21 GPT-5.4 OpenAI90,0 %131128–133 9 03.2026OpenRouter benchmark datasets
22 DeepSeek-V4-Flashoffen DeepSeek89,9 %129116–143 2 zu wenige Daten 04.2026OpenRouter benchmark datasets
23 MiMo-V2.6 Pro Xiaomi Corp89,9 %131117–144 2 zu wenige Daten 09.2026OpenRouter benchmark datasets
24 Tencent Hy4 previewoffen Tencent89,7 %129115–143 2 zu wenige Daten 08.2026OpenRouter benchmark datasets
25 Qwen3.8 Flash Alibaba89,6 %125111–138 2 zu wenige Daten 08.2026OpenRouter benchmark datasets
26 Claude Opus 4.8 Anthropic89,6 %132129–135 8 05.2026OpenRouter benchmark datasets
27 Claude Opus 5 Anthropic89,6 %136133–140 8 07.2026OpenRouter benchmark datasets
28 GPT-5.6 Terra OpenAI89,3 %134131–137 8 07.2026OpenRouter benchmark datasets
29 DeepSeek V4 Pro 0813offen DeepSeek89,3 %128125–131 8 08.2026OpenRouter benchmark datasets
30 Claude Opus 4.7 Anthropic89,2 %127124–130 9 04.2026OpenRouter benchmark datasets
31 Claude Fable 5.1 Anthropic88,9 %140136–143 8 09.2026OpenRouter benchmark datasets
32 Gemini 3 Flash Google DeepMind88,6 %122119–125 9 12.2025OpenRouter benchmark datasets
33 Kimi K2.7 Codeoffen Moonshot88,5 %118115–122 7 06.2026OpenRouter benchmark datasets
34 DeepSeek V4.1 Flashoffen DeepSeek88,4 %128125–131 8 09.2026OpenRouter benchmark datasets
35 GPT-6 Luna OpenAI88,2 %130127–133 8 09.2026OpenRouter benchmark datasets
36 Kimi K2.6offen Moonshot88,0 %121118–124 9 04.2026OpenRouter benchmark datasets
37 GPT-5.2 OpenAI87,9 %125122–128 10 12.2025OpenRouter benchmark datasets
38 Qwen3.7-Plus Alibaba87,4 %115110–119 5 06.2026OpenRouter benchmark datasets
39 Qwen3.5 397B-A17Boffen Alibaba87,3 %114109–118 5 02.2026OpenRouter benchmark datasets
40 GPT-5.6 Luna OpenAI87,2 %130127–133 8 07.2026OpenRouter benchmark datasets
41 Inkling-Smalloffen Thinking Machines87,0 %118115–122 7 07.2026OpenRouter benchmark datasets
42 Claude Haiku 5.5 Anthropic86,9 %129126–132 8 10.2026OpenRouter benchmark datasets
43 GPT-5.1 OpenAI86,9 %120115–125 7 11.2025OpenRouter benchmark datasets
44 MiniMax-M2.1offen MiniMax86,9 %120106–133 2 zu wenige Daten 12.2025OpenRouter benchmark datasets
45 GLM-5.3offen Z.ai (Zhipu AI)86,4 %125122–128 8 08.2026OpenRouter benchmark datasets
46 Claude Fable 5 Anthropic86,2 %138134–141 8 06.2026OpenRouter benchmark datasets
47 Claude Opus 4.5 Anthropic86,1 %115111–118 10 11.2025OpenRouter benchmark datasets
48 Kimi K2.5offen Moonshot85,9 %119112–125 5 01.2026OpenRouter benchmark datasets
49 Claude Sonnet 5 Anthropic85,8 %126123–129 8 06.2026OpenRouter benchmark datasets
50 MiniMax-M2.7offen MiniMax85,7 %120107–134 2 zu wenige Daten 03.2026OpenRouter benchmark datasets
51 Claude Opus 4.6 Anthropic85,7 %123120–126 9 02.2026OpenRouter benchmark datasets
52 MiMo-V2.5-Prooffen Xiaomi Corp85,5 %121108–135 2 zu wenige Daten 04.2026OpenRouter benchmark datasets
53 MiniMax-M2.5offen MiniMax85,4 %118105–132 2 zu wenige Daten 02.2026OpenRouter benchmark datasets
54 DeepSeek-V4-Prooffen DeepSeek85,4 %117113–120 9 04.2026OpenRouter benchmark datasets
55 GPT-5 OpenAI85,0 %121118–124 11 08.2025OpenRouter benchmark datasets
56 GPT-5.4 Mini OpenAI84,0 %117113–120 8 03.2026OpenRouter benchmark datasets
57 Gemini 3.5 Flash-Lite Google DeepMind84,0 %109106–113 7 07.2026OpenRouter benchmark datasets
58 GLM-4.7offen Z.ai (Zhipu AI)83,6 %112107–118 5 12.2025OpenRouter benchmark datasets
59 Qwen 3.8 27Boffen Alibaba83,4 %120106–133 2 zu wenige Daten 08.2026OpenRouter benchmark datasets
60 GLM-5.2offen Z.ai (Zhipu AI)83,3 %121118–124 9 06.2026OpenRouter benchmark datasets
61 Qwen 3.6 35B-A3Boffen Alibaba82,7 %111107–115 5 04.2026OpenRouter benchmark datasets
62 DeepSeek-V3.2-Expoffen DeepSeek82,7 %112103–121 4 zu wenige Daten 09.2025OpenRouter benchmark datasets
63 MiMo-V2.5offen Xiaomi Corp82,4 %117104–131 2 zu wenige Daten 04.2026OpenRouter benchmark datasets
64 Inklingoffen Thinking Machines81,8 %114110–117 7 07.2026OpenRouter benchmark datasets
65 Gemma 4 31B IToffen Google DeepMind81,6 %10498–110 4 zu wenige Daten 04.2026OpenRouter benchmark datasets
66 GLM-5.1offen Z.ai (Zhipu AI)81,2 %116112–120 7 04.2026OpenRouter benchmark datasets
67 Claude Sonnet 4.5 Anthropic81,1 %110107–113 11 09.2025OpenRouter benchmark datasets
68 Qwen3.5-122B-A10Boffen Alibaba80,7 %112103–121 3 zu wenige Daten 02.2026OpenRouter benchmark datasets
69 Gemini 2.5 Pro (Jun 2025) Google DeepMind80,6 %108104–112 7 06.2025OpenRouter benchmark datasets
70 DeepSeek-V3.2offen DeepSeek80,3 %114101–127 2 zu wenige Daten 12.2025OpenRouter benchmark datasets
71 Qwen3.6 27Boffen Alibaba80,1 %115110–119 5 04.2026OpenRouter benchmark datasets
72 GPT-5 mini OpenAI80,1 %115112–118 11 08.2025OpenRouter benchmark datasets
73 Kimi K2 Thinkingoffen Moonshot79,7 %11299–126 2 zu wenige Daten 11.2025OpenRouter benchmark datasets
74 Qwen3-235B-A22B-Thinking (Jul 2025)offen Alibaba79,5 %111105–118 5 07.2025OpenRouter benchmark datasets
75 DeepSeek-V3.1-Terminusoffen DeepSeek79,4 %10289–115 2 zu wenige Daten 09.2025OpenRouter benchmark datasets
76 DeepSeek-R1 (May 2025)offen DeepSeek78,5 %104100–109 4 zu wenige Daten 05.2025OpenRouter benchmark datasets
77 Gemma 4 26B A4Boffen Google DeepMind77,6 %106100–112 3 zu wenige Daten 04.2026OpenRouter benchmark datasets
78 GPT-5.4 Nano OpenAI77,6 %118115–121 8 03.2026OpenRouter benchmark datasets
79 DeepSeek-V3.1offen DeepSeek77,1 %10087–113 2 zu wenige Daten 08.2025OpenRouter benchmark datasets
80 Claude Sonnet 4 Anthropic76,7 %9793–101 5 05.2025OpenRouter benchmark datasets
81 GLM-4.6offen Z.ai (Zhipu AI), Tsinghua University76,5 %117106–128 3 zu wenige Daten 09.2025OpenRouter benchmark datasets
82 MiMo-V2.6 Flash Xiaomi Corp75,9 %11299–125 2 zu wenige Daten 09.2026OpenRouter benchmark datasets
83 gpt-oss-120boffen OpenAI75,8 %10699–112 5 08.2025OpenRouter benchmark datasets
84 qwen3-coder-next —75,8 %10390–116 2 zu wenige Daten 02.2026OpenRouter benchmark datasets
85 Mistral Small 4offen Mistral AI75,3 %9683–108 2 zu wenige Daten 03.2026OpenRouter benchmark datasets
86 Qwen3.5-35B-A3Boffen Alibaba75,0 %106100–112 3 zu wenige Daten 02.2026OpenRouter benchmark datasets
87 Kimi K2 (Sep 2025)offen Moonshot74,7 %10592–118 2 zu wenige Daten 09.2025OpenRouter benchmark datasets
88 Qwen3.5-9Boffen Alibaba74,6 %10195–107 3 zu wenige Daten 02.2026OpenRouter benchmark datasets
89 Qwen3-235B-A22B-Instruct (Jul 2025)offen Alibaba73,9 %9685–107 3 zu wenige Daten 07.2025OpenRouter benchmark datasets
90 Gemini 2.5 Flash (Jun 2025) Google DeepMind73,8 %10291–113 3 zu wenige Daten 06.2025OpenRouter benchmark datasets
91 Claude Haiku 4.5 Anthropic73,1 %10499–108 7 10.2025OpenRouter benchmark datasets
92 GLM-5offen Z.ai (Zhipu AI)72,9 %112105–118 4 zu wenige Daten 02.2026OpenRouter benchmark datasets
93 GPT-5 nano OpenAI68,4 %105102–109 10 08.2025OpenRouter benchmark datasets
94 DeepSeek-V3offen DeepSeek66,2 %7773–82 4 zu wenige Daten 12.2024OpenRouter benchmark datasets
95 GPT-4.1 OpenAI66,1 %9187–94 10 04.2025OpenRouter benchmark datasets
96 Llama 4 Maverickoffen Meta AI64,8 %8177–85 6 04.2025OpenRouter benchmark datasets
97 GPT-4.1 mini OpenAI64,7 %9188–95 10 04.2025OpenRouter benchmark datasets
98 DeepSeek-V3 (Mar 2025)offen DeepSeek64,6 %8581–90 4 zu wenige Daten 03.2025OpenRouter benchmark datasets
99 Claude Sonnet 4.6 Anthropic64,6 %113108–118 7 02.2026OpenRouter benchmark datasets
100 Qwen3-30B-A3B-Instruct (Jul 2025)offen Alibaba64,2 %9186–97 4 zu wenige Daten 07.2025OpenRouter benchmark datasets
101 gpt-oss-20boffen OpenAI63,1 %9589–101 3 zu wenige Daten 08.2025OpenRouter benchmark datasets
102 Qwen3-30B-A3Boffen Alibaba60,4 %9488–100 3 zu wenige Daten 04.2025OpenRouter benchmark datasets
103 Qwen3-Coder-480B-A35Boffen Alibaba60,4 %8572–98 2 zu wenige Daten 07.2025OpenRouter benchmark datasets
104 Qwen3-32Boffen Alibaba55,7 %9589–100 5 04.2025OpenRouter benchmark datasets
105 GPT-4o (May 2024) OpenAI54,2 %7066–75 4 zu wenige Daten 05.2024OpenRouter benchmark datasets
106 Qwen3-14Boffen Alibaba53,3 %9387–99 4 zu wenige Daten 04.2025OpenRouter benchmark datasets
107 GPT-4o (Aug 2024) OpenAI52,8 %7368–77 8 08.2024OpenRouter benchmark datasets
108 GPT-4.1 nano OpenAI49,7 %7672–80 8 04.2025OpenRouter benchmark datasets
109 glm-4.7-flash —49,2 %9488–100 3 zu wenige Daten —OpenRouter benchmark datasets
110 Llama 3.3 70Boffen Meta AI47,5 %6661–71 5 12.2024OpenRouter benchmark datasets
111 Qwen2.5-72Boffen Alibaba46,1 %7268–77 4 zu wenige Daten 09.2024OpenRouter benchmark datasets
112 GPT-4o mini OpenAI43,1 %6864–73 9 07.2024OpenRouter benchmark datasets
113 Gemma 3 12Boffen Google DeepMind37,9 %7366–79 4 zu wenige Daten 03.2025OpenRouter benchmark datasets
114 Qwen2.5-7Boffen Alibaba34,2 %5141–60 3 zu wenige Daten 09.2024OpenRouter benchmark datasets
115 Gemma 3 4Boffen Google DeepMind27,1 %6353–73 3 zu wenige Daten 03.2025OpenRouter benchmark datasets
116 Llama 3.1-8Boffen Meta AI26,9 %5549–60 3 zu wenige Daten 07.2024OpenRouter benchmark datasets
117 Llama 3.2 3Boffen Meta AI17,3 %6142–80 2 zu wenige Daten 09.2024OpenRouter benchmark datasets

Datenquellen: Aider polyglot leaderboard data (Apache-2.0); Arena leaderboard dataset on Hugging Face (CC BY 4.0); Berkeley Function Calling Leaderboard, Gorilla project (Apache-2.0); Epoch AI, ‘Capabilities & benchmarking’ (CC BY 4.0); API list prices via LiteLLM's price table (MIT); each price links to the developer's pricing page; Source: OpenRouter (openrouter.ai), licensed under CC BY 4.0: reuse and republish with attribution to OpenRouter..