Which model leads All-Angles-Bench?

Q: Which model leads All-Angles-Bench?

Step3 VL 10B tops the board at 57.21

All-Angles-Bench Leaderboard — Model Scores

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

The field · 9 measured

42.357.21

9 models measured, most on the official card harness. Step3 VL 10B tops the board at 57.21.

9 measured·Lifecycle Updated Oct 8, 2026Score basis

#◆Vendor$/M in$/M out

Leader — ignites whiteField — fill cools with scoreIndependently verified only0 of 9Bars show 42.3–57.21 · ◆ solid = first-party or better · outlined = arm's-length

01Step3 VL 10B+1 altStepFun57.21———

02HY-Embodied 0.5 MoT-2BTencent55.1———

03GLM-4.6V-Flash (9B)+1 altZ.ai53.24$0.30/M$0.90/M$0.011

04MiMo VL RL 2508 (7B)+1 altXiaomi51.62———

05MiMo Embodied 7BXiaomi49———

06Qwen3 VL 4B (Reasoning)Alibaba46.7———

07Qwen3 VL Thinking (8B)+1 altAlibaba45.88$0.18/M$2.1/M$0.025

08InternVL-3.5 (8B)+1 altOpenGVLab45.29———

09Qwen3 VL 2BAlibaba42.3———

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