Skip to content
VECTOR WIREAI INTELLIGENCE
⌘K
UTC
Models739Deals30d254Policy30d65, 14 pendingSources1,010
Pro Mode — raw scores, one benchmark under the lensofficial card harness
The field · 44 measured
10.3695.6

44 models measured, most on the official card harness. Qwen3 235B A22B tops the board at 95.6.

44 measured·Lifecycle Updated Oct 8, 2026Score basis
#◆Vendor$/M in$/M out
Leader — ignites whiteField — fill cools with scoreBars show 10.36–95.6 · ◆ solid = first-party or better · outlined = arm's-length
01Qwen3 235B A22BAlibaba95.6$0.70/M$2.8/M$0.018
02Qwen3 32BAlibaba93.8$0.16/M$0.64/M$0.004
03DeepSeek-V3DeepSeek91.43P$0.24/M$0.90/M$0.006
04Qwen3 30B A3BAlibaba91$0.20/M$0.80/M$0.005
05MiniMax Text 01MiniMax89.13P———
06Gemini 1.5 ProGoogle85.33P———
07Llama 3.1 Nemotron Instruct 70BNVIDIA85———
08Hunyuan Large InstructTencent81.8———
09Qwen2.5 Instruct 72B+1 altAlibaba81.2$0.47/M$0.49/M$0.006
10GPT-4o+1 altOpenAI79.3$5.0/M$15.0/M$0.126
11Claude 3.5 Sonnet+1 altAnthropic79.2$3.0/M$15.0/M$0.114
12Phi 4 Reasoning PlusMicrosoft79———
13DeepSeek-V2.5+1 altDeepSeek76.2———
14Phi 4Microsoft75.4$0.13/M$0.50/M$0.004
15Gemini 2.0 Flash ExpGoogle72.73P———
16Granite 4.1 30BIBM71.023P———
17Llama 3.1 Instruct 405B+2 altsMeta69.3———
18Granite 4.1 8BIBM68.983P$0.05/M$0.10/M$0.001
19Jamba 1.5 LargeAI2165.4$2.0/M$8.0/M$0.076
20Mistral Small 4Mistral58.3$0.15/M$0.60/M$0.006
21Granite 3.3 8B Instruct+1 altIBM57.56$0.03/M$0.25/M$0.002
22Llama 3.1 70B Instruct+1 altMeta55.7$0.56/M$0.56/M$0.010
23Granite 3.2 8B InstructIBM55.253P———
24Mistral Large 3Mistral55.1$0.50/M$1.5/M$0.018
25Ministral 3 14B+3 altsMistral55.1$0.20/M$0.20/M$0.004
26Qwen3 VL 8B Instruct+3 altsAlibaba52.83P$0.18/M$0.70/M$0.008
27Ministral 3 8B+3 altsMistral50.9$0.15/M$0.15/M$0.003
28Jamba 1.5 MiniAI2146.1$0.20/M$0.40/M$0.007
29Qwen3 VL 4B Instruct+3 altsAlibaba43.83P———
30Gemma 3 12B+3 altsGoogle43.63P$0.05/M$0.15/M$0.002
31Qwen3 14B+3 altsAlibaba42.73P$0.35/M$1.4/M$0.020
32Phi 3.5 MoE InstructMicrosoft37.9———
33Granite 4.1 3BIBM37.83P———
34Granite 3.1 8B InstructIBM37.583P———
35Phi 3.5 Mini InstructMicrosoft37———
36Llama 3.1 8B InstructMeta36.433P$0.02/M$0.04/M$0.001
37Phi 4 Mini InstructMicrosoft32.8———
38Gemma 3 4B+3 altsGoogle31.83P$0.05/M$0.10/M$0.002
39Ministral 3 3B+3 altsMistral30.5$0.10/M$0.10/M$0.003
40Granite 3.3 2B InstructIBM28.863P———
41Granite 3.2 2B InstructIBM24.863P———
42Granite 3.1 2B InstructIBM23.33P———
43DeepSeek-R1-Distill-Llama-8BDeepSeek17.173P$0.05/M$0.05/M$0.003
44DeepSeek-R1-Distill-Qwen-7BDeepSeek10.363P$0.15/M$0.15/M$0.014
◆ 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: 44 models scored · 0 independently verified · 20 vendor cross-reference · 24 vendor-reported · 0 with source disagreement. How these tiers are assigned