MidMarketBench

Model index

Model behaviour through one investment lens.

The board compares model behaviour against investment workflow tasks, not general preference. Every score and rank on this pass is synthetic and illustrative.

RankModelProviderVersionOverallWeights
1GPT-5.5OpenAIgpt-5.583.9Closed
2Claude Opus 4.8Anthropic4.882.8Closed
3Gemini 3.5 FlashGoogle3.5 Flash79.5Closed
4GPT-5.4OpenAIgpt-5.478.3Closed
5Claude Opus 4.7Anthropic4.776.8Closed
6Gemini 3.1 ProGoogle3.1 Pro75.2Closed
7Grok 4.20xAI4.2074.0Closed
8GLM-5.2Z.ai5.273.3Open
9Qwen3.7 MaxAlibaba3.7 Max71.9Closed
10Seed 2.0 ProByteDance2.0 Pro69.6Closed
11DeepSeek V4 ProDeepSeekV4 Pro68.0Open
12GPT-5.2 ProOpenAIgpt-5.2-pro66.8Closed