KI-Anbieter

43 Labore und Anbieter über fünf unabhängige Leaderboards.

Anthropic
Claude Mythos Preview97.4
65 Modelle⌀ 48.2
Moonshot AI
Kimi K383.3
16 Modelle⌀ 52.13 Open Weights
OpenAI
GPT-5.6 Sol80.8
109 Modelle⌀ 41.62 Open Weights
Alibaba
Qwen3.8 Max79.8
90 Modelle⌀ 36.553 Open Weights
Google
Gemini 3.1 Pro Preview74.0
65 Modelle⌀ 42.314 Open Weights
xAI
Grok 4.20 Beta174.0
31 Modelle⌀ 50.0
DeepSeek
Deepseek V4 Pro High Preview73.0
33 Modelle⌀ 45.419 Open Weights
Z.ai
Glm 5.2 Max73.0
23 Modelle⌀ 51.711 Open Weights
Baidu
Ernie 5.0 Preview 120372.0
5 Modelle⌀ 54.6
LongCat
Longcat Flash Chat 2602 Exp72.0
2 Modelle⌀ 71.0
Amazon
Amazon Nova Experimental Chat 26 02 1071.0
25 Modelle⌀ 35.5
NVIDIA
Nvidia Nemotron 3 Ultra 550b A55b Nvfp471.0
17 Modelle⌀ 26.31 Open Weights
Tencent
Hunyuan Hy3 Preview71.0
13 Modelle⌀ 59.4
MiniMax
Minimax M2.570.0
9 Modelle⌀ 41.24 Open Weights
Mistral
Mistral Medium 250870.0
34 Modelle⌀ 29.415 Open Weights
StepFun
Step 3.5 Flash70.0
6 Modelle⌀ 54.5
Meta
Muse Spark 1.169.9
43 Modelle⌀ 38.518 Open Weights
Cohere
Command A 03 202568.0
12 Modelle⌀ 39.9
AI2
Olmo 3.1 32b Instruct67.0
7 Modelle⌀ 54.51 Open Weights
Ling
Ling Flash 2.067.0
5 Modelle⌀ 34.0
IBM
Granite 4.1 8b65.0
7 Modelle⌀ 43.92 Open Weights
Inception
Mercury65.0
2 Modelle⌀ 43.5
Microsoft
Phi 3 Medium 4k Instruct60.0
10 Modelle⌀ 46.62 Open Weights
Sarvam AI
Sarvam 105B50.0
3 Modelle⌀ 49.73 Open Weights
Xiaomi
MiMo V2.5 Pro48.0
4 Modelle⌀ 36.64 Open Weights
Nex AGI
Nex N2 Pro44.4
1 Modell⌀ 44.4
Thinking Machines
Inkling Small41.7
2 Modelle⌀ 36.7
Upstage
Solar Pro441.6
4 Modelle⌀ 17.7
ByteDance
Seed 2.0 Pro40.0
2 Modelle⌀ 29.3
Agnes
Agnes 2 5 Pro Alpha39.7
1 Modell⌀ 39.7
WeRide
Ring 2 6 1t31.3
1 Modell⌀ 31.3
K2
K2 Think V222.0
1 Modell⌀ 22.0
ServiceNow
Apriel V1 5 15b Thinker21.6
2 Modelle⌀ 21.2
North
North Mini Code20.2
1 Modell⌀ 20.2
Trinity
Trinity Large Thinking18.4
1 Modell⌀ 18.4
Celeris
Celeris 112.4
1 Modell⌀ 12.4
Reka
Reka Flash6.0
2 Modelle⌀ 4.9
AI21 Labs
Jamba 1 5 Large4.8
2 Modelle⌀ 3.6
Swiss AI
Apertus 70b Instruct2.0
2 Modelle⌀ 1.5
Async
Async Flash V1 50.0
3 Modelle⌀ 0.0
Cogito
Cogito V2 1 Reasoning0.0
1 Modell⌀ 0.0
Ocat
Ocateve 20.0
1 Modell⌀ 0.0
TII
Falcon 20.0
1 Modell⌀ 0.01 Open Weights