Which model leads MATH (EM)?

Q: Which model leads MATH (EM)?

DeepSeek-V4-Pro-Base tops the board at 64.5

MATH (EM) Leaderboard — Model Scores

Pro Mode — raw scores, one benchmark under the lensofficial card harness

The field · 7 measured

43.464.5

7 models measured, most on the official card harness. DeepSeek-V4-Pro-Base tops the board at 64.5.

7 measured·Lifecycle Updated Oct 8, 2026Score basis

#◆Vendor$/M in$/M out

Leader — ignites whiteField — fill cools with scoreIndependently verified only0 of 7Bars show 43.4–64.5 · ◆ solid = first-party or better · outlined = arm's-length

01DeepSeek-V4-Pro-Base+4 altsDeepSeek64.5———

02DeepSeek-V3DeepSeek61.6$0.24/M$0.90/M$0.009

03DeepSeek-V3.2-Base+2 altsDeepSeek60.5———

04DeepSeek-V4-Flash-Base+3 altsDeepSeek57.4———

05Qwen2.5 Instruct 72BAlibaba54.4$0.47/M$0.49/M$0.009

06Llama 3.1 Instruct 405BMeta49———

07DeepSeek-V2DeepSeek43.4———

◆ solid = independently verified · aggregator attested · vendor attributed — outlined = cross-reference · source-attributed. Dimmed ◆ = confidence inferred. “+N alts” = alternate disclosures in the drawer with why-demoted. Prices per million tokens; $/point = input price ÷ score. Every row opens its model dossier.

Verification: 7 models scored · 0 independently verified · 2 vendor cross-reference · 5 vendor-reported · 0 with source disagreement. How these tiers are assigned