Which model leads MMStar (Acc)?

Q: Which model leads MMStar (Acc)?

Qwen2.5 VL 72B tops the board at 70.8

MMStar (Acc) Leaderboard — 8 Models Ranked

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

The field · 8 measured

54.870.8

8 models measured, most on the third-party eval harness. Qwen2.5 VL 72B tops the board at 70.8.

8 measured·Lifecycle Updated Oct 8, 2026Score basis

#◆Vendor$/M in$/M out

Leader — ignites whiteField — fill cools with scoreIndependently verified only0 of 8Bars show 54.8–70.8 · ◆ solid = first-party or better · outlined = arm's-length

01Qwen2.5 VL 72BAlibaba70.8———

02Qwen2.5 VL 32B InstructAlibaba69.5———

03GPT-4o+1 altOpenAI64.7$5.0/M$15.0/M$0.155

04Kimi VL A3B Thinking+1 altMoonshot64.2———

05Gemma 3 27BGoogle63.1$0.08/M$0.16/M$0.002

06Kimi VL A3B Instruct+1 altMoonshot61.3VENDOR———

07Gemma 3 12B+1 altGoogle56.1$0.05/M$0.15/M$0.002

08GPT-4o miniOpenAI54.8$0.15/M$0.60/M$0.007

◆ 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: 8 models scored · 0 independently verified · 6 vendor cross-reference · 2 vendor-reported · 1 with source disagreement. How these tiers are assigned