Which model leads Minerva Math?

Q: Which model leads Minerva Math?

Nemotron 3 Super 120B A12B tops the board at 84.84

Minerva Math Leaderboard — Model Scores

Pro Mode — raw scores, one benchmark under the lensthird-party eval harness

The field · 11 measured

3884.84

11 models measured, most on the third-party eval harness. Nemotron 3 Super 120B A12B tops the board at 84.84.

11 measured·Lifecycle Updated Oct 8, 2026Score basis

#◆Vendor$/M in$/M out

Leader — ignites whiteField — fill cools with scoreIndependently verified only0 of 11Bars show 38–84.84 · ◆ solid = first-party or better · outlined = arm's-length

01Nemotron 3 Super 120B A12BNVIDIA84.84$0.30/M$0.90/M$0.007

02Nemotron 3.5 LightningNVIDIA82.78$0.06/M$0.20/M$0.002

03NVIDIA Nemotron 3 Nano 30B A3BNVIDIA82.64$0.05/M$0.20/M$0.002

04Granite 4.1 30BIBM81.32———

05Granite 4.1 8BIBM80.1$0.05/M$0.10/M$0.001

06Granite 4.1 3BIBM67.94———

07Qwen3.5 35B A3BAlibaba59.66$0.25/M$2.0/M$0.019

08Granite 4.1 30B Base+1 altIBM45.66———

09Gemma 4 26B A4BGoogle43.74$0.07/M$0.34/M$0.005

10Granite 4.1 8B Base+1 altIBM43.42———

11Granite 4.1 3B BaseIBM38———

◆ 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: 11 models scored · 0 independently verified · 2 vendor cross-reference · 9 vendor-reported · 0 with source disagreement. How these tiers are assigned