Which model leads AttaQ?

Q: Which model leads AttaQ?

Granite 3.3 8B Instruct tops the board at 88.5

AttaQ Leaderboard — Model Scores

Pro Mode — raw scores, one benchmark under the lensthird-party eval harness

The field · 12 measured

42.4588.5

12 models measured, most on the third-party eval harness. Granite 3.3 8B Instruct tops the board at 88.5.

12 measured·Lifecycle Updated Oct 8, 2026Score basis

#◆Vendor$/M in$/M out

Leader — ignites whiteField — fill cools with scoreIndependently verified only0 of 12Bars show 42.45–88.5 · ◆ solid = first-party or better · outlined = arm's-length

01Granite 3.3 8B Instruct+1 altIBM88.5VENDOR$0.03/M$0.25/M$0.002

02Granite 3.3 2B InstructIBM87.47———

03Granite 4.1 30BIBM85.76———

04Granite 3.1 8B InstructIBM85.73———

05Granite 3.2 8B InstructIBM84.7———

06Granite 3.1 2B InstructIBM84.7———

07Llama 3.1 8B InstructMeta83.43$0.02/M$0.04/M$0.000

08Granite 3.2 2B InstructIBM83.23———

09Granite 4.1 3BIBM81.88———

10Granite 4.1 8BIBM81.19$0.05/M$0.10/M$0.001

11DeepSeek-R1-Distill-Llama-8BDeepSeek42.87$0.05/M$0.05/M$0.001

12DeepSeek-R1-Distill-Qwen-7BDeepSeek42.45$0.15/M$0.15/M$0.004

◆ 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: 12 models scored · 0 independently verified · 8 vendor cross-reference · 4 vendor-reported · 0 with source disagreement. How these tiers are assigned