Skip to content
VECTOR WIREAI INTELLIGENCE
⌘K
UTC
Models741Deals30d262Policy30d64, 11 pendingSources1,015
Pro Mode — raw scores, one benchmark under the lensofficial card harness
The field · 107 measured
4.0892.5

107 models measured, most on the official card harness. DeepSeek-V4-Pro tops the board at 92.5.

107 measured·15 new this week·Lifecycle Updated Oct 9, 2026Score basis
#◆Vendor$/M in$/M out
Leader — ignites whiteField — fill cools with scoreBars show 4.08–92.5 · ◆ solid = first-party or better · outlined = arm's-length
01DeepSeek-V4-ProDeepSeek92.53P$0.43/M$0.87/M$0.007
02Qwen3.8 Flash Next+1 altAlibaba91.9$0.15/M$0.47/M$0.003
03Gemini 3.1 Pro+1 altGoogle91.73P$2.0/M$12.0/M$0.076
04Qwen3.7 MaxAlibaba91.6$2.5/M$7.5/M$0.055
05DeepSeek-V4-Flash+1 altDeepSeek90.6$0.44/M$1.3/M$0.010
06Namazusakana90.33$0.95/M$4.0/M$0.027
07Qwen3.8 27B+2 altsAlibaba90.3$0.50/M$3.0/M$0.019
08Qwen3.7 Plus Preview+2 altsAlibaba89.6$0.40/M$1.6/M$0.011
09Kimi K2.6+2 altsMoonshot89.6$0.95/M$4.0/M$0.028
10Nemotron 3 Ultra 550B A55B+1 altNVIDIA89$0.60/M$2.5/M$0.017
11Claude Opus 4.6+2 altsAnthropic88.8$5.0/M$25.0/M$0.169
12Seed 2.0ByteDance87.8$0.50/M$3.0/M$0.020
13MAI-Thinking-1+1 altMicrosoft87.7———
14Gemini 3 Pro+2 altsGoogle87.43P$2.0/M$12.0/M$0.080
15Qwen3.6 PlusAlibaba87.1$0.50/M$3.0/M$0.020
16GPT-5.1OpenAI873P$1.3/M$10.0/M$0.065
17GPT-5+2 altsOpenAI87$1.3/M$10.0/M$0.065
18Step 3.5 Flash+3 altsStepFun86.4$0.10/M$0.30/M$0.002
19Qwen3 Max (Reasoning)Alibaba85.9$1.2/M$6.0/M$0.042
20GLM-5.1Z.ai85.73P$1.4/M$4.4/M$0.034
21Kimi K2.5+5 altsMoonshot85$0.45/M$2.3/M$0.016
22GLM-4.7+4 altsZ.ai84.9$0.60/M$2.2/M$0.016
23Claude Opus 4.5+1 altAnthropic84.83P$5.0/M$25.0/M$0.177
24Qwen3.6 27B+2 altsAlibaba83.9$0.60/M$3.6/M$0.025
25Qwen3.5 397B A17B+2 altsAlibaba83.6$0.60/M$3.6/M$0.025
26DeepSeek-V3.2+6 altsDeepSeek83.3$0.28/M$0.42/M$0.004
27Kimi K2 Thinking+5 altsMoonshot83.1$0.60/M$2.5/M$0.019
28Ling-3.0-flashInclusionAI82.8$0.07/M$0.22/M$0.002
29GLM-4.6+1 altZ.ai82.8$0.57/M$2.2/M$0.017
30EXAONE 4.5 33BLG AI81.4———
31Qwen3.5 27B+2 altsAlibaba80.7$0.30/M$2.4/M$0.017
32MiMo V2 Flash+4 altsXiaomi80.6$0.10/M$0.30/M$0.002
33Qwen3.6 35B A3B+2 altsAlibaba80.4$0.38/M$2.3/M$0.016
34Gemma 4 31BNEW+29 altsGoogle80$0.17/M$0.40/M$0.004
35Seed 1.8ByteDance79.5$0.25/M$2.0/M$0.014
36Qwen3.5 122B A10BAlibaba78.9$0.40/M$3.2/M$0.023
37MiniMax M2.7MiniMax77.23P$0.21/M$0.84/M$0.007
38Gemma 4 26B A4BNEW+32 altsGoogle77.1$0.07/M$0.34/M$0.003
39Granite 4.2 30B+1 altIBM75.77$0.16/M$0.65/M$0.005
40Qwen3.5 35B A3B+1 altAlibaba74.6$0.25/M$2.0/M$0.015
41Granite 4.2 8B+2 altsIBM73.24$0.06/M$0.25/M$0.002
42Gemma 4 12B+23 altsGoogle72$0.10/M$0.30/M$0.003
43Sarvam 105BSarvam AI71.7$0.04/M$0.17/M$0.001
44North-Mini-Code-1.0Cohere70.3———
45Qwen3 VL 235B A22B ReasoningAlibaba70.1$0.40/M$4.0/M$0.031
46Sarvam 30BSarvam AI70$0.03/M$0.11/M$0.001
47Granite 4.2 3B+2 altsIBM69.71$0.03/M$0.12/M$0.001
48Qwen3 Next 80B A3BAlibaba68.7$0.15/M$1.2/M$0.010
49NVIDIA Nemotron 3 Nano 30B A3BNVIDIA68.3$0.05/M$0.20/M$0.002
50Seed Oss 36B InstructByteDance67.43P$0.21/M$0.57/M$0.006
51Qwen3 30B A3B 2507 Thinking+1 altAlibaba663P$0.20/M$2.4/M$0.020
52Qwen3 VL 32B ReasoningAlibaba65.6$0.16/M$0.64/M$0.006
53Qwen3.5 9BAlibaba65.6$0.14/M$0.20/M$0.003
54Qwen3 VL 30B A3B ReasoningAlibaba64.2$0.20/M$2.4/M$0.020
55GLM-4.7-FlashZ.ai643P$0.06/M$0.40/M$0.004
56Claude Sonnet 4.5+2 altsAnthropic64$3.0/M$15.0/M$0.141
57GPT Oss 20bOpenAI613P$0.07/M$0.18/M$0.002
58LFM2.5-2.6BLiquid AI59.41———
59Qwen3 VL Thinking (8B)Alibaba58.6$0.18/M$2.1/M$0.019
60Qwen3 Next 80B A3B InstructAlibaba56.6$0.15/M$1.2/M$0.012
61Qwen3.5 4BAlibaba55.8$0.03/M$0.15/M$0.002
62Qwen3 VL 235B A22B InstructAlibaba54.3$0.40/M$1.6/M$0.018
63Kimi K2 Instruct+3 altsMoonshot53.7$0.57/M$2.3/M$0.027
64Gemma 4 E4BNEW+31 altsGoogle52$0.02/M$0.10/M$0.001
65Qwen3 235B A22B Instruct 2507Alibaba51.8$0.23/M$0.92/M$0.011
66Qwen3 VL 4B (Reasoning)Alibaba51.3———
67MiMo 7B RL+8 altsXiaomi49.3———
68Qwen3 4B 2507 InstructNEWAlibaba48.723P$0.01/M$0.03/M$0.000
69Claude Sonnet 4+1 altAnthropic48.53P$3.0/M$15.0/M$0.186
70Claude Opus 4+1 altAnthropic47.43P$15.0/M$75.0/M$0.949
71DeepSeek-V3+1 altDeepSeek46.93P$0.24/M$0.90/M$0.012
72O1 Mini+4 altsOpenAI46.8———
73MiMo 7B SFT+4 altsXiaomi45.5———
74GPT-4.1+1 altOpenAI44.73P$2.0/M$8.0/M$0.112
75Gemma 4 E2BNEW+37 altsGoogle44———
76Qwen3 VL 32B InstructAlibaba43.8$0.16/M$0.64/M$0.009
77MiMo 7B RL Zero+4 altsXiaomi42.9———
78Qwen3 VL 30B A3B InstructAlibaba42.6$0.20/M$0.80/M$0.012
79MiMo V2.5 ProXiaomi39.6$0.43/M$0.87/M$0.016
80MiMo V2.5 Pro Base+2 altsXiaomi39.6———
81Qwen3 VL 8B InstructAlibaba39.3$0.18/M$0.70/M$0.011
82QwQ 32B Preview+4 altsAlibaba39.1$0.66/M$1.0/M$0.021
83Qwen3 VL 4B InstructAlibaba37.9———
84Claude 3.5 Sonnet+4 altsAnthropic37.2$3.0/M$15.0/M$0.242
85Qwen3 235B A22B+1 altAlibaba373P$0.70/M$2.8/M$0.047
86Qwen3 4BNEWAlibaba36.113P———
87MiMo V2.5 Base+2 altsXiaomi35.5———
88Hy3-preview-Base+1 altTencent34.86———
89DeepSeek-R1-Distill-Qwen-14B+4 altsDeepSeek31.9$0.20/M$0.20/M$0.006
90GPT-4o+4 altsOpenAI30.9$5.0/M$15.0/M$0.324
91MiMo V2 Flash BaseXiaomi30.83P———
92DeepSeek-V3-Base+1 altDeepSeek29.31———
93MiMo 7B Base+4 altsXiaomi29.1———
94Gemma 3 27BNEW+27 altsGoogle29.1$0.08/M$0.16/M$0.004
95GLM-4.5-Base+1 altZ.ai27.43———
96Granite 4.0 H TinyNEWIBM26.733P———
97Kimi K2 Base+6 altsMoonshot26.3———
98DeepSeek-V3.2-Exp-BaseDeepSeek24.93P———
99DeepSeek-V3.1-BaseDeepSeek24.83P———
100Qwen3 1.7BNEWAlibaba24.073P———
101DeepSeek-R1-Distill-Qwen-7B+4 altsDeepSeek23.9$0.15/M$0.15/M$0.006
102LFM2-8B-A1B (Non-Reasoning)NEWLiquid AI21.043P———
103Gemma 3 4BNEWGoogle18.863P$0.05/M$0.10/M$0.004
104LFM2-2.6BNEWLiquid AI14.413P———
105Llama 3.2 3B InstructNEWMeta11.473P$0.05/M$0.33/M$0.017
106Gemma 3 1B InstructNEWGoogle4.273P———
107Llama 3.2 Instruct 1BNEWMeta4.083P$0.03/M$0.20/M$0.028
◆ 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: 107 models scored · 0 independently verified · 38 vendor cross-reference · 69 vendor-reported · 0 with source disagreement. How these tiers are assigned