Which model leads AGI EVAL?

Q: Which model leads AGI EVAL?

Granite 4.1 30B tops the board at 77.8

AGI EVAL Leaderboard — 13 Models Ranked

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

The field · 13 measured

28.577.8

13 models measured, most on the third-party eval harness. Granite 4.1 30B tops the board at 77.8.

13 measured·Lifecycle Updated Oct 8, 2026Score basis

#◆Vendor$/M in$/M out

Leader — ignites whiteField — fill cools with scoreIndependently verified only0 of 13Bars show 28.5–77.8 · ◆ solid = first-party or better · outlined = arm's-length

01Granite 4.1 30BIBM77.8———

02Granite 4.1 8BIBM72.43$0.05/M$0.10/M$0.001

03Granite 4.1 30B Base+1 altIBM69.2———

04Granite 4.1 3BIBM65.16———

05Granite 4.1 8B Base+1 altIBM61.68———

06Gemma 2 27B ITGoogle55.1VENDOR$0.65/M$0.65/M$0.012

07Granite 4.1 3B BaseIBM54.32———

08Gemma 2 Instruct (9B)Google52.8VENDOR———

09Gemma (7B)Google42.1———

10Llama 3 Instruct 8BMeta42VENDOR$0.04/M$0.14/M$0.002

11Phi 3 Mini 4K InstructMicrosoft39———

12Mistral 7BMistral35.1———

13ERNIE 4.5Baidu28.5VENDOR———

◆ 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: 13 models scored · 0 independently verified · 3 vendor cross-reference · 10 vendor-reported · 0 with source disagreement. How these tiers are assigned