Skip to content
VECTOR WIREAI INTELLIGENCE
⌘K
UTC
Models741Deals30d270Policy30d65, 13 pendingSources1,015
Pro Mode — raw scores, one benchmark under the lensofficial card harness
The field · 28 measured
56.287.6

28 models measured, most on the official card harness. Gemini 3 Pro tops the board at 87.6.

28 measured·Lifecycle Updated Oct 8, 2026Score basis
#◆Vendor$/M in$/M out
Leader — ignites whiteField — fill cools with scoreBars show 56.2–87.6 · ◆ solid = first-party or better · outlined = arm's-length
01Gemini 3 Pro+1 altGoogle87.6$2.0/M$12.0/M$0.080
02Gemini 3 Flash PreviewGoogle86.9$0.50/M$3.0/M$0.020
03Kimi K2.5+1 altMoonshot86.6$0.45/M$2.3/M$0.016
04GPT-5.2+1 altOpenAI85.9$1.8/M$14.0/M$0.092
05Qwen3.7 Plus PreviewAlibaba85.4$0.40/M$1.6/M$0.012
06Gemini 3.1 Flash Lite PreviewGoogle84.8$0.25/M$1.5/M$0.010
07GPT-5OpenAI84.6$1.3/M$10.0/M$0.066
08MiniMax M3MiniMax84.6$0.28/M$1.1/M$0.008
09Qwen3.6 27BAlibaba84.4$0.60/M$3.6/M$0.025
10Claude Opus 4.5Anthropic84.43P$5.0/M$25.0/M$0.178
11Qwen3.6 PlusAlibaba84$0.50/M$3.0/M$0.021
12Qwen3.6 35B A3BAlibaba83.7$0.38/M$2.3/M$0.016
13Gemini 2.5 Pro+1 altGoogle83.6$1.3/M$10.0/M$0.067
14O3OpenAI83.3$2.0/M$8.0/M$0.060
15Seed 1.8ByteDance82.7$0.25/M$2.0/M$0.014
16Qwen3.5 27BAlibaba82.3$0.30/M$2.4/M$0.016
17Qwen3.5 122B A10BAlibaba82$0.40/M$3.2/M$0.022
18Qwen3.5 35B A3BAlibaba80.4$0.25/M$2.0/M$0.014
19Qwen3 VL 235B A22B Reasoning+1 altAlibaba80$0.40/M$4.0/M$0.028
20Qwen3 VL 32B ReasoningAlibaba79$0.16/M$0.64/M$0.005
21Qwen3 VL 30B A3B ReasoningAlibaba75$0.20/M$2.4/M$0.017
22Qwen3 VL 235B A22B InstructAlibaba74.7$0.40/M$1.6/M$0.013
23Qwen3 VL Thinking (8B)Alibaba72.8$0.18/M$2.1/M$0.016
24Qwen3 VL 4B (Reasoning)Alibaba69.4———
25Qwen3 VL 30B A3B InstructAlibaba68.7$0.20/M$0.80/M$0.007
26Qwen3 VL 8B InstructAlibaba65.3$0.18/M$0.70/M$0.007
27GPT-4oOpenAI61.2$5.0/M$15.0/M$0.163
28Qwen3 VL 4B InstructAlibaba56.2———
◆ 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: 28 models scored · 0 independently verified · 1 vendor cross-reference · 27 vendor-reported · 0 with source disagreement. How these tiers are assigned