Skip to content
VECTOR WIREAI INTELLIGENCE
⌘K
UTC
Models741Deals30d261Policy30d65, 14 pendingSources1,013
Pro Mode — raw scores, one benchmark under the lensofficial card harness
The field · 23 measured
8.550

23 models measured, most on the official card harness. Claude Sonnet 4.5 tops the board at 50.

23 measured·Lifecycle Updated Oct 8, 2026Score basis
#◆Vendor$/M in$/M out
Leader — ignites whiteField — fill cools with scoreBars show 8.5–50 · ◆ solid = first-party or better · outlined = arm's-length
01Claude Sonnet 4.5+1 altAnthropic50$3.0/M$15.0/M$0.180
02MiniMax M2.1MiniMax47.9$0.30/M$1.2/M$0.016
03MiniMax M2+1 altMiniMax46.3$0.30/M$1.2/M$0.016
04GPT-5OpenAI43.83P$1.3/M$10.0/M$0.128
05Claude Opus 4.1Anthropic43.3$15.0/M$75.0/M$1.039
06Claude Opus 4+3 altsAnthropic43.2$15.0/M$75.0/M$1.042
07Claude Haiku 4.5Anthropic41$1.0/M$5.0/M$0.073
08GLM-4.6+1 altZ.ai40.5$0.57/M$2.2/M$0.034
09DeepSeek-V3.2DeepSeek37.73P$0.28/M$0.42/M$0.009
10DeepSeek-V3.2-ExpDeepSeek37.7$0.28/M$0.42/M$0.009
11GLM-4.5Z.ai37.5$0.60/M$2.2/M$0.037
12Claude Sonnet 4+4 altsAnthropic35.5$3.0/M$15.0/M$0.254
13Claude 3.7 SonnetAnthropic35.2$3.0/M$15.0/M$0.256
14LongCat Flash LiteMeituan33.75———
15GLM-4.7Z.ai33.3$0.60/M$2.2/M$0.042
16DeepSeek-V3.1DeepSeek31.3$0.56/M$1.7/M$0.036
17MiMo V2 FlashXiaomi30.5$0.10/M$0.30/M$0.007
18GPT-4.1+1 altOpenAI30.33P$2.0/M$8.0/M$0.165
19Kimi K2 Instruct+4 altsMoonshot30$0.57/M$2.3/M$0.048
20GLM-4.5-AirZ.ai30$0.17/M$0.98/M$0.019
21Nemotron 3 Super 120B A12BNVIDIA25.78$0.30/M$0.90/M$0.023
22Gemini 2.5 ProGoogle25.33P$1.3/M$10.0/M$0.222
23NVIDIA Nemotron 3 Nano 30B A3BNVIDIA8.5$0.05/M$0.20/M$0.015
◆ 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: 23 models scored · 0 independently verified · 4 vendor cross-reference · 19 vendor-reported · 1 with source disagreement. How these tiers are assigned