Fresh evidence
New results
Completed benchmark runs, newest first. Every entry links to a permanent static result page.
Published runs
202 canonical results across 46 models.
- Gemini 3.5 Flash-Lite · GermEval NER (DE)85.4%
Google Micro-F1 (no MISC) provider-internal
- Gemini 3.5 Flash-Lite · INCLUDE (DE)85.0%
Google Exact match provider-internal
- Gemini 3.5 Flash-Lite · MuSR (DE)82.3%
Google Accuracy provider-internal
- Gemini 3.5 Flash-Lite · SB10K (DE)62.5%
Google MCC provider-internal
- Gemini 3.5 Flash-Lite · ScaLA (DE)75.1%
Google MCC provider-internal
- GLM-5.3 Flash · GermEval NER (DE)86.4%
Z.ai Micro-F1 (no MISC) fp8
- GLM-5.3 Flash · INCLUDE (DE)86.9%
Z.ai Exact match fp8
- GLM-5.3 Flash · MuSR (DE)81.9%
Z.ai Accuracy fp8
- GLM-5.3 Flash · SB10K (DE)63.6%
Z.ai MCC fp8
- GLM-5.3 Flash · ScaLA (DE)77.2%
Z.ai MCC fp8
- Tencent HY3 · GermEval NER (DE)76.1%
Tencent Micro-F1 (no MISC) fp8
- Tencent HY3 · INCLUDE (DE)83.2%
Tencent Exact match fp8
- Tencent HY3 · MuSR (DE)85.0%
Tencent Accuracy
- Tencent HY3 · SB10K (DE)62.4%
Tencent MCC fp8
- Tencent HY3 · ScaLA (DE)69.7%
Tencent MCC fp8
- Tencent HY4-Preview · GermEval NER (DE)82.1%
Tencent Micro-F1 (no MISC) fp8
- Tencent HY4-Preview · INCLUDE (DE)86.9%
Tencent Exact match fp8
- Tencent HY4-Preview · MuSR (DE)80.1%
Tencent Accuracy fp8
- 62.4%
- 74.0%
- Gemini 3.1 Pro · MMLU-ProX (DE)89.2%
Google Exact match provider-internal
- Muse Glimmer 30B · INCLUDE (DE)86.0%
Meta Exact match
- Muse Glimmer 30B · MuSR (DE)86.1%
Meta Accuracy
- 86.9%
- Qwen3.6 27B (BF16) · GermEval NER (DE)82.7%
Alibaba Micro-F1 (no MISC) bf16
- Qwen3.6 27B (BF16) · INCLUDE (DE)83.2%
Alibaba Exact match bf16
- Qwen3.6 27B (BF16) · MMLU-ProX (DE)82.3%
Alibaba Exact match bf16
- Qwen3.6 27B (BF16) · MuSR (DE)84.4%
Alibaba Accuracy bf16
- Qwen3.6 27B (BF16) · SB10K (DE)61.0%
Alibaba MCC bf16
- Qwen3.6 27B (BF16) · ScaLA (DE)71.7%
Alibaba MCC bf16
- Qwen3.8 27B (BF16) · GermEval NER (DE)81.5%
Alibaba Micro-F1 (no MISC) bf16
- Qwen3.8 27B (BF16) · INCLUDE (DE)87.9%
Alibaba Exact match bf16
- Qwen3.8 27B (BF16) · MMLU-ProX (DE)79.9%
Alibaba Exact match bf16
- Qwen3.8 27B (BF16) · MuSR (DE)86.1%
Alibaba Accuracy bf16
- Qwen3.8 27B (BF16) · SB10K (DE)62.6%
Alibaba MCC bf16
- Qwen3.8 27B (BF16) · ScaLA (DE)66.9%
Alibaba MCC bf16
- Qwen3.8 27B (FP8) · GermEval NER (DE)81.6%
Alibaba Micro-F1 (no MISC) fp8
- Qwen3.8 27B (FP8) · INCLUDE (DE)86.9%
Alibaba Exact match fp8
- Qwen3.8 27B (FP8) · MMLU-ProX (DE)80.0%
Alibaba Exact match fp8
- 62.7%
- 66.3%
- DeepSeek V4 Pro 0813 · MMLU-ProX (DE)80.7%
DeepSeek Exact match
- Gemini 3.7 Flash · GermEval NER (DE)86.9%
Google Micro-F1 (no MISC) provider-internal
- Gemini 3.7 Flash · INCLUDE (DE)93.5%
Google Exact match provider-internal
- Gemini 3.7 Flash · MMLU-ProX (DE)88.2%
Google Exact match provider-internal
- Gemini 3.7 Flash · MuSR (DE)88.6%
Google Accuracy provider-internal
- Gemini 3.7 Flash · SB10K (DE)67.5%
Google MCC provider-internal
- Gemini 3.7 Flash · ScaLA (DE)79.1%
Google MCC provider-internal
- DeepSeek V4 Pro 0813 · GermEval NER (DE)81.7%
DeepSeek Micro-F1 (no MISC) provider-internal
- DeepSeek V4 Pro 0813 · INCLUDE (DE)87.9%
DeepSeek Exact match provider-internal
- DeepSeek V4 Pro 0813 · MuSR (DE)82.3%
DeepSeek Accuracy provider-internal
- DeepSeek V4 Pro 0813 · SB10K (DE)60.8%
DeepSeek MCC provider-internal
- DeepSeek V4 Pro 0813 · ScaLA (DE)75.6%
DeepSeek MCC provider-internal
- GPT-5.6 Luna · GermEval NER (DE)81.9%
OpenAI Micro-F1 (no MISC) provider-internal
- GPT-5.6 Luna · INCLUDE (DE)86.0%
OpenAI Exact match provider-internal
- GPT-5.6 Luna · MuSR (DE)75.8%
OpenAI Accuracy provider-internal
- GPT-5.6 Luna · SB10K (DE)61.9%
OpenAI MCC provider-internal
- GPT-5.6 Luna · ScaLA (DE)72.2%
OpenAI MCC provider-internal
- GPT-5.6 Terra · INCLUDE (DE)88.8%
OpenAI Exact match provider-internal
- GPT-5.6 Terra · MuSR (DE)78.7%
OpenAI Accuracy provider-internal
- GPT-5.6 Terra · GermEval NER (DE)83.3%
OpenAI Micro-F1 (no MISC) provider-internal
- GPT-5.6 Terra · SB10K (DE)62.7%
OpenAI MCC provider-internal
- GPT-5.6 Terra · ScaLA (DE)75.1%
OpenAI MCC provider-internal
- NVIDIA Nemotron 3 Ultra 550B-A55B · GermEval NER (DE)80.3%
NVIDIA Micro-F1 (no MISC)
- DeepSeek V4 Flash 0423 · GermEval NER (DE)80.7%
DeepSeek Micro-F1 (no MISC) fp4
- DeepSeek V4 Flash 0423 · MuSR (DE)85.3%
DeepSeek Accuracy fp4
- 61.9%
- 74.0%
- DeepSeek V4 Flash 0731 · GermEval NER (DE)82.0%
DeepSeek Micro-F1 (no MISC) fp4
- 61.2%
- 76.1%
- MiMo V2.5 · GermEval NER (DE)80.8%
Xiaomi Micro-F1 (no MISC) fp8
- MiMo V2.5 · SB10K (DE)61.2%
Xiaomi MCC fp8
- MiMo V2.5 · ScaLA (DE)53.4%
Xiaomi MCC fp8
- Gemini 3.5 Flash-Lite · MMLU-ProX (DE)80.2%
Google Exact match provider-internal
- Gemini 3.6 Flash · GermEval NER (DE)86.2%
Google Micro-F1 (no MISC) provider-internal
- Gemini 3.6 Flash · INCLUDE (DE)74.8%
Google Exact match provider-internal
- Gemini 3.6 Flash · MMLU-ProX (DE)86.0%
Google Exact match provider-internal
- Gemini 3.6 Flash · MuSR (DE)80.3%
Google Accuracy provider-internal
- Gemini 3.6 Flash · SB10K (DE)64.5%
Google MCC provider-internal
- Gemini 3.6 Flash · ScaLA (DE)77.6%
Google MCC provider-internal
- Claude Sonnet 5 · INCLUDE (DE)72.7%
Anthropic Exact match provider-internal
- Claude Fable 5 · GermEval NER (DE)87.7%
Anthropic Micro-F1 (no MISC) provider-internal
- Claude Fable 5 · SB10K (DE)67.3%
Anthropic MCC provider-internal
- Claude Fable 5 · ScaLA (DE)83.6%
Anthropic MCC provider-internal
- Claude Fable 5 · INCLUDE (DE)72.7%
Anthropic Exact match provider-internal
- Claude Fable 5 · MuSR (DE)89.5%
Anthropic Accuracy provider-internal
- GPT-5.5 · MMLU-ProX (DE)87.2%
OpenAI Exact match provider-internal
- Gemini 3 Flash Preview · GermEval NER (DE)84.7%
Google Micro-F1 (no MISC) provider-internal
- Gemini 3 Flash Preview · INCLUDE (DE)71.9%
Google Exact match provider-internal
- Gemini 3 Flash Preview · MMLU-ProX (DE)86.3%
Google Exact match provider-internal
- Gemini 3 Flash Preview · MuSR (DE)79.3%
Google Accuracy provider-internal
- Gemini 3 Flash Preview · SB10K (DE)63.6%
Google MCC provider-internal
- Gemini 3 Flash Preview · ScaLA (DE)77.0%
Google MCC provider-internal
- DeepSeek V4 Pro · MuSR (DE)85.3%
DeepSeek Accuracy provider-internal
- DeepSeek V4 Flash · MMLU-ProX (DE)74.9%
DeepSeek Exact match provider-internal
- DeepSeek V4 Pro · MMLU-ProX (DE)80.8%
DeepSeek Exact match provider-internal
- Claude Opus 4.8 · GermEval NER (DE)85.8%
Anthropic Micro-F1 (no MISC) provider-internal
- Claude Opus 4.8 · MuSR (DE)86.3%
Anthropic Accuracy provider-internal
- Claude Opus 4.8 · SB10K (DE)64.5%
Anthropic MCC provider-internal
- Claude Opus 4.8 · ScaLA (DE)82.6%
Anthropic MCC provider-internal
- GPT-5.5 · GermEval NER (DE)85.9%
OpenAI Micro-F1 (no MISC) provider-internal
- GPT-5.5 · MuSR (DE)86.3%
OpenAI Accuracy provider-internal
- GPT-5.5 · SB10K (DE)62.5%
OpenAI MCC provider-internal
- GPT-5.5 · ScaLA (DE)79.7%
OpenAI MCC provider-internal
- Claude Opus 4.8 · INCLUDE (DE)71.9%
Anthropic Exact match provider-internal
- GPT-5.5 · INCLUDE (DE)74.8%
OpenAI Exact match provider-internal
- Gemma 4 12B · GermEval NER (DE)79.6%
Google Micro-F1 (no MISC) bf16
- Gemma 4 12B · INCLUDE (DE)69.1%
Google Exact match bf16
- Gemma 4 12B · MMLU-ProX (DE)73.1%
Google Exact match bf16
- Gemma 4 12B · MuSR (DE)81.6%
Google Accuracy bf16
- Gemma 4 12B · SB10K (DE)62.2%
Google MCC bf16
- Gemma 4 12B · ScaLA (DE)64.1%
Google MCC bf16
- Gemma 4 31B · GermEval NER (DE)82.6%
Google Micro-F1 (no MISC) bf16
- Gemma 4 31B · INCLUDE (DE)68.3%
Google Exact match bf16
- Gemma 4 31B · MMLU-ProX (DE)82.1%
Google Exact match bf16
- Gemma 4 31B · MuSR (DE)83.5%
Google Accuracy bf16
- Gemma 4 31B · SB10K (DE)61.0%
Google MCC bf16
- Gemma 4 31B · ScaLA (DE)71.0%
Google MCC bf16
- Qwen3 14B · GermEval NER (DE)73.0%
Alibaba Micro-F1 (no MISC) bf16
- Qwen3 14B · INCLUDE (DE)63.3%
Alibaba Exact match bf16
- Qwen3 14B · MMLU-ProX (DE)64.1%
Alibaba Exact match bf16
- Qwen3 14B · MuSR (DE)69.7%
Alibaba Accuracy bf16
- Qwen3 14B · SB10K (DE)56.1%
Alibaba MCC bf16
- Qwen3 14B · ScaLA (DE)64.0%
Alibaba MCC bf16
- Qwen3.5 9B · GermEval NER (DE)72.6%
Alibaba Micro-F1 (no MISC) bf16
- Qwen3.5 9B · INCLUDE (DE)64.7%
Alibaba Exact match bf16
- Qwen3.5 9B · MMLU-ProX (DE)73.4%
Alibaba Exact match bf16
- Qwen3.5 9B · MuSR (DE)80.9%
Alibaba Accuracy bf16
- Qwen3.5 9B · SB10K (DE)55.7%
Alibaba MCC bf16
- DeepSeek V4 Flash · MuSR (DE)83.5%
DeepSeek Accuracy fp8
- Gemini 3.1 Flash-Lite · INCLUDE (DE)71.9%
Google Exact match provider-internal
- Gemini 3.1 Pro · INCLUDE (DE)77.7%
Google Exact match provider-internal
- Gemini 3.5 Flash · INCLUDE (DE)74.1%
Google Exact match provider-internal
- Gemma 4 26B A4B · MuSR (DE)83.7%
Google Accuracy bf16
- GLM-5.1 · MuSR (DE)81.4%
Z.ai Accuracy fp8
- MiMo V2.5 Pro · MuSR (DE)80.9%
Xiaomi Accuracy fp8
- Qwen3.5 9B · ScaLA (DE)62.4%
Alibaba MCC bf16
- DeepSeek V4 Flash · GermEval NER (DE)80.2%
DeepSeek Micro-F1 (no MISC) bf16
- DeepSeek V4 Flash · SB10K (DE)62.4%
DeepSeek MCC bf16
- DeepSeek V4 Flash · ScaLA (DE)74.9%
DeepSeek MCC bf16
- DeepSeek V4 Pro · GermEval NER (DE)82.2%
DeepSeek Micro-F1 (no MISC) fp8
- DeepSeek V4 Pro · SB10K (DE)60.0%
DeepSeek MCC fp8
- DeepSeek V4 Pro · ScaLA (DE)73.5%
DeepSeek MCC fp8
- Gemini 2.5 Flash · GermEval NER (DE)81.8%
Google Micro-F1 (no MISC) provider-internal
- Gemini 2.5 Flash · MuSR (DE)83.3%
Google Accuracy provider-internal
- Gemini 2.5 Flash · SB10K (DE)61.6%
Google MCC provider-internal
- Gemini 2.5 Flash · ScaLA (DE)77.1%
Google MCC provider-internal
- Gemini 3.1 Flash-Lite · MuSR (DE)84.4%
Google Accuracy provider-internal
- Gemini 3.1 Pro · GermEval NER (DE)87.3%
Google Micro-F1 (no MISC) provider-internal
- Gemini 3.1 Pro · MuSR (DE)88.1%
Google Accuracy provider-internal
- Gemini 3.1 Pro · SB10K (DE)70.1%
Google MCC provider-internal
- Gemini 3.1 Pro · ScaLA (DE)80.4%
Google MCC provider-internal
- Gemini 3.5 Flash · GermEval NER (DE)86.0%
Google Micro-F1 (no MISC) provider-internal
- Gemini 3.5 Flash · MuSR (DE)84.2%
Google Accuracy provider-internal
- Gemini 3.5 Flash · SB10K (DE)64.3%
Google MCC provider-internal
- Gemini 3.5 Flash · ScaLA (DE)79.4%
Google MCC provider-internal
- GLM-5.1 · GermEval NER (DE)52.2%
Z.ai Micro-F1 (no MISC) fp8
- GLM-5.1 · SB10K (DE)23.1%
Z.ai MCC fp8
- GLM-5.1 · ScaLA (DE)69.9%
Z.ai MCC fp8
- MiMo V2.5 Pro · GermEval NER (DE)81.9%
Xiaomi Micro-F1 (no MISC) fp8
- MiMo V2.5 Pro · SB10K (DE)61.7%
Xiaomi MCC fp8
- MiMo V2.5 Pro · ScaLA (DE)74.8%
Xiaomi MCC fp8
- Qwen3.6 35B-A3B · GermEval NER (DE)79.9%
Alibaba Micro-F1 (no MISC) fp8
- Qwen3.6 35B-A3B · SB10K (DE)58.4%
Alibaba MCC fp8
- Qwen3.6 35B-A3B · ScaLA (DE)64.2%
Alibaba MCC fp8
- Qwen3.7 Max · GermEval NER (DE)82.1%
Alibaba Micro-F1 (no MISC) provider-internal
- Qwen3.7 Max · SB10K (DE)63.2%
Alibaba MCC provider-internal
- Qwen3.7 Max · ScaLA (DE)69.8%
Alibaba MCC provider-internal
- Tencent HY3-Preview · GermEval NER (DE)77.3%
Tencent Micro-F1 (no MISC) provider-internal
- Tencent HY3-Preview · SB10K (DE)38.9%
Tencent MCC provider-internal
- Tencent HY3-Preview · ScaLA (DE)62.5%
Tencent MCC provider-internal
- DeepSeek V4 Flash · INCLUDE (DE)70.5%
DeepSeek Exact match fp8
- DeepSeek V4 Pro · INCLUDE (DE)70.5%
DeepSeek Exact match fp8
- Gemini 3.1 Flash-Lite · GermEval NER (DE)83.2%
Google Micro-F1 (no MISC) provider-internal
- Gemini 3.1 Flash-Lite · SB10K (DE)60.4%
Google MCC provider-internal
- Gemini 3.1 Flash-Lite · ScaLA (DE)74.2%
Google MCC provider-internal
- Gemma 4 26B A4B · GermEval NER (DE)81.8%
Google Micro-F1 (no MISC) provider-internal
- Gemma 4 26B A4B · SB10K (DE)15.2%
Google MCC provider-internal
- Gemma 4 26B A4B · ScaLA (DE)66.3%
Google MCC provider-internal
- GLM-5.1 · INCLUDE (DE)82.2%
Z.ai Exact match fp8
- Kimi K2.6 · INCLUDE (DE)69.8%
Moonshot Exact match fp8
- MiMo V2.5 Pro · INCLUDE (DE)67.6%
Xiaomi Exact match fp8
- Qwen3.7 Max · INCLUDE (DE)71.2%
Alibaba Exact match provider-internal
- Gemma 4 26B A4B · MMLU-ProX (DE)78.2%
Google Exact match bf16
- Claude Haiku 4.5 · MMLU-ProX (DE)75.3%
Anthropic Exact match provider-internal
- Gemini 2.5 Flash-Lite · MMLU-ProX (DE)71.2%
Google Exact match provider-internal
- Gemini 3.1 Flash-Lite · MMLU-ProX (DE)82.2%
Google Exact match provider-internal
- Gemma 3 12B · INCLUDE (DE)53.2%
Google Exact match
- Gemma 4 26B A4B · INCLUDE (DE)64.7%
Google Exact match bf16
- Grok 4.3 · INCLUDE (DE)63.3%
xAI Exact match
- Ministral 14B · INCLUDE (DE)58.3%
Mistral Exact match
- Claude Haiku 4.5 · INCLUDE (DE)68.3%
Anthropic Exact match provider-internal
- Gemini 2.5 Flash · INCLUDE (DE)70.5%
Google Exact match provider-internal
- Qwen3.6 35B-A3B · INCLUDE (DE)68.3%
Alibaba Exact match fp8
- Tencent HY3-Preview · INCLUDE (DE)69.1%
Tencent Exact match provider-internal
- Qwen3.6 35B-A3B · MMLU-ProX (DE)80.0%
Alibaba Exact match fp8
- gpt-oss-120b · INCLUDE (DE)66.2%
OpenAI Exact match
- Tencent HY3-Preview · MMLU-ProX (DE)73.2%
Tencent Exact match provider-internal
- Gemini 2.5 Flash · MMLU-ProX (DE)79.6%
Google Exact match provider-internal
- Gemini 3.5 Flash · MMLU-ProX (DE)86.5%
Google Exact match provider-internal
- Undated MiniMax M2.7 · INCLUDE (DE)72.7%
MiniMax Exact match fp8