deepseek-ai/DeepSeek-V4-Pro
last-exam-pass-rate · 0 source-defined score; unit unspecified
Source-qualified evaluations, report statistics and safety observations.
last-exam-pass-rate · 0 source-defined score; unit unspecified
table · 65.86 source-defined score; unit unspecified
prompt_level_loose · 77 source-defined score; unit unspecified
mmlu_pro · 56.7 source-defined score; unit unspecified
ru · 64.2 source-defined score; unit unspecified
diamond · 81.3131313131 source-defined score; unit unspecified
skew · 75.24 source-defined score; unit unspecified
MathArena/aime_2026 · 95.7 source-defined score; unit unspecified
swe_bench_multilingual_%_resolved · 36.47 source-defined score; unit unspecified
pass-rate-strict · 9.18 source-defined score; unit unspecified
pt · 77.4 source-defined score; unit unspecified
taubench_airline · 0.12 source-defined score; unit unspecified
en · 82.2 source-defined score; unit unspecified
overall · 32.81 source-defined score; unit unspecified
precision · 99.27 source-defined score; unit unspecified
hle · 25.3 source-defined score; unit unspecified
recall · 14.28 source-defined score; unit unspecified
fruitloops_windows · 35.1 source-defined score; unit unspecified
These charts contain model-linked results. The charts above also include unlinked source profiles. Compare scores within each evaluation. Model quality, usage, transparency and system performance measure different things.