Skip to content
VECTOR WIREAI INTELLIGENCE
⌘K
UTC
Models741Deals30d262Policy30d65, 14 pendingSources1,013
Pro Mode — raw scores, one benchmark under the lensofficial card harness
The field · 38 measured
32.6780.6

38 models measured, most on the official card harness. Qwen3 VL 235B A22B Reasoning tops the board at 80.6.

38 measured·3 new this week·Lifecycle Updated Oct 8, 2026Score basis
#◆Vendor$/M in$/M out
Leader — ignites whiteField — fill cools with scoreBars show 32.67–80.6 · ◆ solid = first-party or better · outlined = arm's-length
01Qwen3 VL 235B A22B ReasoningAlibaba80.6$0.40/M$4.0/M$0.027
02Qwen3 VL 235B A22B InstructAlibaba78.7$0.40/M$1.6/M$0.013
03Qwen3 VL 32B ReasoningAlibaba78.1$0.16/M$0.64/M$0.005
04Claude Sonnet 4.5Anthropic77.8$3.0/M$15.0/M$0.116
05Qwen3 VL 32B InstructAlibaba76$0.16/M$0.64/M$0.005
06Qwen3 VL 30B A3B ReasoningAlibaba76$0.20/M$2.4/M$0.017
07Qwen3 VL 30B A3B InstructAlibaba74.2$0.20/M$0.80/M$0.007
08Qwen3 VL Thinking (8B)Alibaba74.1$0.18/M$2.1/M$0.015
09Qwen3 VL 4B (Reasoning)Alibaba70.8———
10Qwen3 VL 8B InstructAlibaba69.6$0.18/M$0.70/M$0.006
11GPT-4oOpenAI69.1$5.0/M$15.0/M$0.145
12Claude 3.5 Sonnet+1 altAnthropic68.3$3.0/M$15.0/M$0.132
13Qwen3 VL 4B InstructAlibaba67.4———
14Qwen2.5 VL 72B+1 altAlibaba65.783P———
15Gemma 3 27BGoogle64.9$0.08/M$0.16/M$0.002
16Qwen2 VL 72B Instruct+1 altAlibaba64.5———
17InternVL3_5-4BOpenGVLab60.7———
18Gemma 3 12BGoogle59.6$0.05/M$0.15/M$0.002
19Gemini 1.5 ProGoogle54.11———
20GPT-4o miniOpenAI52.1$0.15/M$0.60/M$0.007
21InternVL3_5-2BOpenGVLab52———
22Qwen3.5 4BAlibaba50.3$0.03/M$0.15/M$0.002
23Gemma 4 E4BGoogle49.3$0.02/M$0.10/M$0.001
24Gemma 3 4BGoogle48.8$0.05/M$0.10/M$0.002
25InternVL3-2BNEWOpenGVLab48.73P———
26LFM2.5-VL-3B+1 altLiquid AI48.4———
27InternVL-2-8BOpenGVLab46.33———
28LFM2-VL-3BLiquid AI45.6———
29InternVL-2-4BOpenGVLab44.22———
30Qwen3.5 2BAlibaba44.1———
31Phi 3.5 Vision InstructMicrosoft43———
32InternVL3.5-1BOpenGVLab41.893P———
33Gemma 4 E2BGoogle41.1———
34LFM2.5-VL-1.6BLiquid AI40.563P———
35LFM2-VL-1.6BNEW+1 altLiquid AI39.673P———
36LFM2-VL-450MNEW+1 altLiquid AI34.443P———
37North-Micro-Vision-InstructCohere32.9———
38LFM2.5-VL-450M-ExtractLiquid AI32.673P———
◆ 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: 38 models scored · 0 independently verified · 10 vendor cross-reference · 28 vendor-reported · 0 with source disagreement. How these tiers are assigned