Skip to content
VECTOR WIREAI INTELLIGENCE
UTC
Pro Mode — raw scores, one benchmark under the lensofficial card harness
The field · 21 measured
39.9385.5

21 models measured, most on the official card harness. Qwen2 VL 72B Instruct tops the board at 85.5.

21 measured·Lifecycle Updated Oct 8, 2026Score basis
#◆Vendor$/M in$/M out
Leader — ignites whiteField — fill cools with scoreBars show 39.93–85.5 · ◆ solid = first-party or better · outlined = arm's-length
01Qwen2 VL 72B InstructAlibaba85.5———
02LFM2.5-VL-3BLiquid AI84.3———
03Qwen2.5 VL 72B+1 altAlibaba83.263P———
04LFM2-VL-3BLiquid AI83———
05Qwen3.5 4BAlibaba81.2$0.03/M$0.15/M$0.001
06Qwen3 VL 4B (Reasoning)Alibaba80.553P———
07Qwen2.5 VL 3BAlibaba79.093P———
08InternVL3_5-4BOpenGVLab77.5———
09Qwen3.5 2BAlibaba77.3———
10Mage-VL-4BMicrosoft77.283P———
11InternVL3_5-2BOpenGVLab76.6———
12Phi 4 R V 15BMicrosoft76.063P———
13Phi 3.5 Vision InstructMicrosoft72———
14GPT-4o miniOpenAI70.9$0.15/M$0.60/M$0.005
15Claude 3.5 SonnetAnthropic70.5$3.0/M$15.0/M$0.128
16Gemma 4 E4BGoogle69$0.02/M$0.10/M$0.001
17InternVL-2-8BOpenGVLab68.8———
18InternVL-2-4BOpenGVLab66.2———
19Gemini 1.5 ProGoogle64.5———
20Gemma 4 E2BGoogle62.5———
21Phi 4 Multimodal InstructMicrosoft39.933P———
◆ 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: 21 models scored · 0 independently verified · 8 vendor cross-reference · 13 vendor-reported · 0 with source disagreement. How these tiers are assigned