‹ PublicAI Index
The LLM benchmark aggregator.
Grok 4.5
SpaceXAI
Strongest in Secure code (#2 of 270), weakest in Toxicity avoidance (#189 of 268). Above par in 26 of 29 scopes. Among the models it meets almost everywhere, it finishes behind Claude Fable 5 and Claude Fable 5.1 and ahead of GPT-6 Sol and O3.
§ 1 · Profile
What it is good at
Bars run from 50 — the average of the models each source lists — so right of the line is above par. The middle column is the gap to whoever leads that scope.
Core abilities59.3−6#15/1992/4
General intelligence65.1−4.3#10/1781/3
Language56.2−10.4#21/621/1
Instruction following56.2−17.6#22/621/1
Knowledge54.2−8.3#44/1971/3
Academic knowledge60−3.9#14/1231/1
Human preference63−5.8#46/3541/1
Human preference63−5.8#46/3541/1
Safety55−6.4#52/3291/3
Secure code66−0.2#2/2701/1
Fairness57.4−12.6#47/2921/2
Harm refusal52.9−8.6#110/2921/2
Jailbreak resistance52.9−12.1#143/2681/1
Toxicity avoidance50.8−8.2#189/2681/1
Coding53.3−12.9#53/1793/6
Agentic coding53.5−13.4#54/1733/5
Professional53.2−9.2#57/1941/2
Cybersecurity55.4−8.5#21/471/1
IT operations39.6−34.2#29/331/1
Legal55.8−10.1#38/1731/1
Medical54.4−12.1#49/1491/1
Finance54.2−8.2#62/1621/1
Reasoning53.3−12.1#61/1954/5
Science61.8−1.7#12/1221/1
Reasoning56.9−12.1#41/1472/3
Mathematics46.6−17.2#100/1522/2
Agents51.7−14.7#105/2983/6
Web research38.1−25.8#31/351/2
Knowledge work59.7−11#49/2082/2
§ 2 · Head to head
What it beats, and what beats it
The same models turn up scope after scope. Counted once: where both were placed, who finished higher. Bars run right for Grok 4.5, left for the other.
§ 3 · Sources
Where the numbers come from
10 publications, 38 figures. Every one links to the page it was read from.
LMArena Text 1466
GDPval-AA 1389
AA-Briefcase 1278
Terminal-Bench 12.4% (Grok Build)
ARC-AGI-2 52.6%
LiveBench 1.0 75.8LiveBench 1.0 · Reasoning 87.2LiveBench 1.0 · Coding 68.6LiveBench 1.0 · Agentic Coding 56.5LiveBench 1.0 · Mathematics 90.8LiveBench 1.0 · Data Analysis 73LiveBench 1.0 · Language 82.8LiveBench 1.0 · Instruction Following 71.5
Kagi LLM Benchmark 83.5%
SimpleBench 70%
Vals · Legal Research Bench 37.98%Vals · LegalBench 85.97%Vals · Finance Agent 48.35%Vals · CorpFin 67.41%Vals · TaxEval 71.67%Vals · MortgageTax 61.8%Vals · MedCode 43.29%Vals · MedScribe 86.88%Vals · CyberBench 71.01%Vals · SRE Bench 0.76%Vals · Web Search Index 38.75% (Exa)Vals · SWE-bench Verified 86.6% (Mini-SWE-agent)Vals · Vibe Code Bench 69% (OpenHands)Vals · Code Migration 36.6%Vals · GPQA Diamond 92.93%Vals · MMLU Pro 89.22%Vals · ProofBench 31%
Enkrypt · Jailbreak risk 10.4%Enkrypt · Harmful content risk 0%Enkrypt · CBRN risk 25.3%Enkrypt · Toxicity risk 5.6%Enkrypt · Bias risk 66.4%Enkrypt · Insecure code risk 0.4%
Badge
[](https://publicai.io/model-index/m/grok-4-5)