Model · xAI

xAI Grok 4.5

4 appearances · meta-score 72.5/100 · last seen 2026-07-22 · scored under rubric v15

Pooled composite across 4 primary runs: 62.8 ± 2.5 (95% CI)

$103.11Total cost
151,511,297Tokens
85.4%Cache hit
2,858Requests
0.14%Failure rate
1.18Tools/response
1,469,402Tokens per $

Agent role scores

Per-agent-role scores for xAI Grok 4.5
Agent role	Score / 10
Sprint-Review Agent	8.71
Orchestrator	8.14
Develop Agent	8.11
Build Agent	7.80
Sprint-Plan Agent	7.58
Scout Agent	7.36
Spec Agent	7.19
Code-Review Agent	7.02
Architect	6.99
Code Agent	6.75
Platform Agent	6.31
Test Agent	6.22
Pseudocode Agent	5.56

Run history (primary model)

Runs where xAI Grok 4.5 was the primary model, oldest first
Run	SoW	Date	Composite
Grok 4.5	circles	Jul 2026	61.0
Grok 4.5	circles	Jul 2026	63.2
Grok 4.5	circles	Jul 2026	62.3
Grok 4.5	circles	Jul 2026	64.7

Every run on Favur Evals is scored by the same deterministic engine, on the same Statements of Work. The benchmark is self-funded — no vendor sponsorship, credits, or grants.