Which model leads OSWorld-G?

Q: Which model leads OSWorld-G?

Qwen3 VL 235B A22B Reasoning tops the board at 68.3

OSWorld-G Leaderboard — 6 Models Ranked

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

The field · 6 measured

31.9168.3

6 models measured, most on the official card harness. Qwen3 VL 235B A22B Reasoning tops the board at 68.3.

6 measured·Lifecycle Updated Oct 8, 2026Score basis

#◆Vendor$/M in$/M out

Leader — ignites whiteField — fill cools with scoreIndependently verified only0 of 6Bars show 31.91–68.3 · ◆ solid = first-party or better · outlined = arm's-length

01Qwen3 VL 235B A22B ReasoningAlibaba68.3$0.40/M$4.0/M$0.032

02Step3 VL 10B+1 altStepFun59.02———

03Qwen3 VL Thinking (8B)+1 altAlibaba56.7$0.18/M$2.1/M$0.020

04GLM-4.6V-Flash (9B)+1 altZ.ai54.71$0.30/M$0.90/M$0.011

05MiMo VL RL 2508 (7B)+1 altXiaomi50.54———

06InternVL-3.5 (8B)+1 altOpenGVLab31.91———

◆ 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: 6 models scored · 0 independently verified · 4 vendor cross-reference · 2 vendor-reported · 0 with source disagreement. How these tiers are assigned