Skip to content
VECTOR WIREAI INTELLIGENCE
⌘K
UTC
Models741Deals30d269Policy30d65, 14 pendingSources1,014
Pro Mode — raw scores, one benchmark under the lensofficial card harness
The field · 64 measured
4.4598.8

64 models measured, most on the official card harness. Gemini 2.5 Pro tops the board at 98.8.

64 measured·6 new this week·Lifecycle Updated Oct 8, 2026Score basis
#◆Vendor$/M in$/M out
Leader — ignites whiteField — fill cools with scoreBars show 4.45–98.8 · ◆ solid = first-party or better · outlined = arm's-length
01Gemini 2.5 Pro+1 altGoogle98.83P$1.3/M$10.0/M$0.057
02Sarvam 105BSarvam AI98.6$0.04/M$0.17/M$0.001
03GLM-4.5Z.ai98.2$0.60/M$2.2/M$0.014
04O3+1 altOpenAI98.13P$2.0/M$8.0/M$0.051
05GLM-4.5-AirZ.ai98.1$0.17/M$0.98/M$0.006
06NVIDIA Nemotron Nano 9B V2NVIDIA97.8$0.04/M$0.16/M$0.001
07Kimi K2 Instruct+3 altsMoonshot97.4$0.57/M$2.3/M$0.015
08DeepSeek-R1+7 altsDeepSeek97.3$2.0/M$4.0/M$0.031
09Sarvam 30BSarvam AI97$0.03/M$0.11/M$0.001
10Llama 3.1 Nemotron Ultra 253B V1NVIDIA97———
11MiniMax M1 80K+2 altsMiniMax96.8$0.55/M$2.2/M$0.014
12Qwen3 14BAlibaba96.8$0.35/M$1.4/M$0.009
13LongCat Flash LiteMeituan96.8———
14O1+4 altsOpenAI96.4$15.0/M$60.0/M$0.389
15Claude 3.7 SonnetAnthropic96.2$3.0/M$15.0/M$0.094
16MiniMax M1 40K+2 altsMiniMax96———
17DeepSeek-R1-Zero+1 altDeepSeek95.9———
18MiMo 7B RL+3 altsXiaomi95.83P———
19Llama 3.1 Nemotron Nano 8B V1NVIDIA95.4———
20DeepSeek-R1-Distill-Llama-70B+4 altsDeepSeek94.5$0.70/M$1.1/M$0.010
21Claude Opus 4+3 altsAnthropic94.43P$15.0/M$75.0/M$0.477
22DeepSeek-R1-Distill-Qwen-32B+4 altsDeepSeek94.3$0.29/M$0.29/M$0.003
23Claude Sonnet 4+1 altAnthropic943P$3.0/M$15.0/M$0.096
24DeepSeek-R1-Distill-Qwen-14B+9 altsDeepSeek93.9$0.20/M$0.20/M$0.002
25MiMo 7B RL ZeroXiaomi93.63P———
26DeepSeek-R1-Distill-Qwen-7B+9 altsDeepSeek92.8$0.15/M$0.15/M$0.002
27Qwen3 235B A22B+3 altsAlibaba91.23P$0.70/M$2.8/M$0.019
28QwQ 32B Preview+9 altsAlibaba90.6$0.66/M$1.0/M$0.009
29DeepSeek-V3+9 altsDeepSeek90.2$0.24/M$0.90/M$0.006
30O1 Mini+10 altsOpenAI90———
31DeepSeek-R1-Distill-Llama-8B+4 altsDeepSeek89.1$0.05/M$0.05/M$0.001
32LFM2.5-1.2B-ThinkingLiquid AI87.963P———
33Qwen3 4B 2507 InstructNEWAlibaba85.63P$0.01/M$0.03/M$0.000
34DeepSeek-R1-Distill-Qwen-1.5B+4 altsDeepSeek83.9———
35Qwen3 1.7BAlibaba81.923P———
36Qwen3.5 2BAlibaba81.52———
37Qwen2.5 Instruct 72BAlibaba80$0.47/M$0.49/M$0.006
38Claude 3.5 Sonnet+10 altsAnthropic78.3$3.0/M$15.0/M$0.115
39DeepSeek-V2.5DeepSeek74.7———
40GPT-4o+10 altsOpenAI74.6$5.0/M$15.0/M$0.134
41LFM2-8B-A1B (Non-Reasoning)NEWLiquid AI74.23P———
42Llama 3.1 Instruct 405BMeta73.8———
43Gemma 3 4BNEWGoogle73.23P$0.05/M$0.10/M$0.001
44SynLogic 7BMiniMax71.83P———
45Qwen3 1.7B (instruct)Alibaba70.43P———
46Granite 3.3 8B Instruct+1 altIBM69.02$0.03/M$0.25/M$0.002
47Qwen2.5 7B BaseAlibaba64.63P———
48LFM2-2.6BNEWLiquid AI63.63P———
49Granite 4.0 H TinyNEWIBM58.23P———
50Granite 3.3 2B InstructIBM58.093P———
51Qwen3.5 0.8BAlibaba57.6———
52DeepSeek-V2DeepSeek56.3———
53Granite 3.2 8B InstructIBM52.83P———
54Granite 3.1 8B InstructIBM48.733P———
55Granite 4.0 H 1BIBM47.23P———
56Gemma 3 1B InstructGoogle45.23P———
57Granite 4.0 1BIBM44.83P———
58Llama 3.2 3B InstructNEWMeta41.23P$0.05/M$0.33/M$0.005
59Granite 3.2 2B InstructIBM35.543P———
60Granite 3.1 2B InstructIBM35.073P———
61Llama 3.2 Instruct 1BMeta23.43P$0.03/M$0.20/M$0.005
62LFM2.5-8B-A1B-DSparkLiquid AI8.023P———
63LFM2.5-1.2B-Instruct+1 altLiquid AI5.783P———
64LFM2.5-2.6B-DSparkLiquid AI4.453P———
◆ 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: 64 models scored · 0 independently verified · 27 vendor cross-reference · 37 vendor-reported · 0 with source disagreement. How these tiers are assigned