This benchmark uses all publicly recorded configurations. Rankings are grouped by metric and evaluation protocol. Recorded tool access, budgets, and frameworks are shown separately. Scores remain available for reference when conditions are unspecified.
Exact Match
Tool access not recorded / Reasoning budget not recorded / HAI-LLM · Official DeepSeek-V3 Technical Report Table 3; internal HAI-LLM framework; all listed models share the same evaluation setting.
MMMLU · DeepSeek-V3 Technical Report Table 3; MMMLU non-English (EM); unit is not printed · Exact Match · Tool access not recorded / Reasoning budget not recorded / HAI-LLM · Official DeepSeek-V3 Technical Report Table 3; internal HAI-LLM framework; all listed models share the same evaluation setting.