KI-Rangliste

43 Modelle benchmarked · aktualisiert am 28. Juni 2026

Agentic Coding

🥇Claude Fable 5 (Adaptive Reasoning, Max Effort, Opus 4.8 Fallback)Anthropic76.5
🥈GPT-5.5 (xhigh)OpenAI74.9
🥉Gemini 3.5 Flash (high)Google70.1
4Qwen3.7 MaxAlibaba66.0
5DeepSeek V4 Pro (Reasoning, Max Effort)DeepSeek59.4
6Muse SparkMeta58.6
7Grok Build 0.1 0616xAI51.5
8Nemotron 3 Ultra 550B A55B (Reasoning)NVIDIA49.3
9Mistral Medium 3.5Mistral46.9

Konversation

🥇Claude Fable 5 (Adaptive Reasoning, Max Effort, Opus 4.8 Fallback)Anthropic59.9
🥈GPT-5.5 (xhigh)OpenAI54.8
🥉Gemini 3.5 Flash (high)Google50.2
4Qwen3.7 MaxAlibaba46.0
5DeepSeek V4 Pro (Reasoning, Max Effort)DeepSeek44.3
6Muse SparkMeta43.1
7Grok Build 0.1 0616xAI39.8
8Nemotron 3 Ultra 550B A55B (Reasoning)NVIDIA37.8
9Mistral Medium 3.5Mistral29.9

Bildbearbeitung

🥇GPT Image 1.5 (high)OpenAI1261
🥈Nano Banana 2 (Gemini 3.1 Flash Image Preview)Google1249
🥉grok-imagine-image-qualityxAI1235
4Wan 2.7 ProAlibaba1204

Bild zu Video

🥇grok-imagine-videoxAI1325
🥈Veo 3.1 Fast PreviewGoogle1278
🥉Wan 2.7Alibaba1266
4Cosmos3-Super-Image2VideoNVIDIA1251
5SoraOpenAI971

Text zu Bild

🥇GPT Image 2 (high)OpenAI1338
🥈Nano Banana 2 (Gemini 3.1 Flash Image Preview)Google1254
🥉Cosmos3-Super-Text2Image (agentic)NVIDIA1225
4grok-imagine-image-qualityxAI1201
5Qwen Image 2.0 Pro (2026-04-22)Alibaba1174
6Janus ProDeepSeek712

Sprachsynthese

🥇Gemini 3.1 Flash TTSGoogle1216
🥈Fun-Realtime-TTSAlibaba1210
🥉xAI Text to SpeechxAI1197
4TTS-1 HDOpenAI1097
5Voxtral TTSMistral1075
6Magpie-Multilingual 357M (Feb 2026)NVIDIA1056

Text zu Video

🥇grok-imagine-videoxAI1233
🥈Veo 3 PreviewGoogle1225
🥉Wan 2.7Alibaba1221
4Sora 2 ProOpenAI1191

Spoton orchestriert das alles für Sie

Warum sich entscheiden, wenn Sie das Beste jeder KI für jede Aufgabe haben können?