Grok 4.5

Grok 4.5 is a reasoning model from xAI. 43 benchmarks count toward its score, in 7 categories.

availableShows if the model has enough results for an index.
IndexOverall score out of 100.66.0 ±3.0
CoverageShare of the index weight with results.95%
SpeedOutput tokens per second.49/s
Input / 1MUS dollars per 1M input tokens.$2
Output / 1MUS dollars per 1M output tokens.$6
ContextMaximum tokens in one request.500K
EloLMArena rating and rank.1450 (#45)

The index is a score out of 100. The ± range shows how much it can change.

30,103 votes. Elo shows what people prefer. It does not change the score.

CapabilitiesScore per category, out of 100.

Out of 100
AgenticMulti-step tasks with tools.
64.1
CodingCode writing and repair.
68.5
ReasoningLogic problems and puzzles.
67.5
MultimodalTasks with images and text.
46.6
KnowledgeFacts and expert knowledge.
68.2
MultilingualTasks in many languages.
N/A
InstructionTasks with strict rules in the prompt.
72.9
MathMath problems.
64.1

Results

43 counted
BenchmarkThe test name.CategoryThe capability that the test measures.ResultThe score from the publisher.IndexThis result as a score out of 100.RunThe settings of the run.DateDate of the result.Published byThe source of the result.
OTIS Mock AIME 2024-2025Math97.8%66.2high effortEpoch AI
GPQA diamondKnowledge93.4%64.5high effortEpoch AI
Artificial Analysis GPQA DiamondKnowledge93.1%64.0Artificial Analysis
GPQA DiamondKnowledge92.9%64.01 Sept 2026Vals AI
LiveBench MathematicsMath90.8%69.325 Jun 2026LiveBench
VulcanBench v3Coding89.9%77.4VulcanBench contributors
MMLU ProKnowledge89.2%61.11 Sept 2026Vals AI
LiveCodeBenchCoding87.4%63.71 Sept 2026Vals AI
LiveBench ReasoningReasoning87.2%77.125 Jun 2026LiveBench
SWE-benchCoding86.6%67.01 Sept 2026Vals AI
ARC-AGI-1 (semi-private)Reasoning85.7%67.7high effortARC Prize Foundation
LiveBench LanguageKnowledge82.8%70.425 Jun 2026LiveBench
Artificial Analysis MMMU-ProMultimodal80.4%65.3Artificial Analysis
Artificial Analysis Long Context ReasoningReasoning79.3%63.1Artificial Analysis
LiveBench Data AnalysisReasoning73.0%57.425 Jun 2026LiveBench
Artificial Analysis Coding IndexCoding72.5%70.0Artificial Analysis
LiveBench Instruction FollowingInstruction71.5%72.925 Jun 2026LiveBench
Vibe Code Bench v1.1Coding69.0%70.9OpenHands21 Sept 2026Vals AI
LiveBench CodingCoding68.6%51.625 Jun 2026LiveBench
Terminal-Bench 2.1Agentic67.8%64.121 Sept 2026Vals AI
cursorBench32Coding66.7%74.8Benchmark authors
SkillsBenchCoding66.0%79.0OpenHands11 Sept 2026Vals AI
SWE-bench ProCoding64.7%66.6Xiang Deng et al.
MMMU ProMultimodal61.8%27.91 Sept 2026Vals AI
FrontierMath-Tiers-1-3-v2-PrivateMath57.2%63.6high effortEpoch AI
LiveBench Agentic CodingAgentic56.5%69.925 Jun 2026LiveBench
Artificial Analysis SciCodeCoding55.0%69.2Artificial Analysis
DeepSWEAgentic53.0%62.3Datacurve AI
ARC-AGI-2 (semi-private)Reasoning52.6%66.3high effortARC Prize Foundation
Artificial Analysis Omniscience AccuracyKnowledge51.6%77.7Artificial Analysis
SimpleQA VerifiedKnowledge48.3%66.3high effortEpoch AI
τ²-bench BankingAgentic47.9%33.2high effort · Sierra4 Aug 2026Sierra Research
GDPval-AA normalizedAgentic43.5%69.0Artificial Analysis
Artificial Analysis Humanity's Last ExamKnowledge42.7%70.9Artificial Analysis
Artificial Analysis Agentic IndexAgentic42.1%71.5Artificial Analysis
IOICoding40.6%63.321 Sept 2026Vals AI
Artificial Analysis Intelligence IndexKnowledge38.8%70.9Artificial Analysis
Code MigrationCoding36.6%68.321 Sept 2026Vals AI
Chess PuzzlesReasoning36.0%70.1high effortEpoch AI
ProofBench v1.1Math31.0%62.021 Sept 2026Vals AI
FrontierMath-Tier-4-v2-PrivateMath24.4%59.6high effortEpoch AI
Terminal-Bench 3.0Agentic15.7%66.8Ryan Marten et al.
Critical Physics TasksReasoning15.4%70.7Artificial Analysis
Terminal-Bench 4.0.0Agentic12.4%68.2high effort · Grok Build21 Sept 2026Terminal-Bench
Agent Arena command recoveryAgentic6.674.815 Sept 2026LMArena
Terminal-Bench 4.0Agentic6.6%64.121 Sept 2026Vals AI
Agent Arena steerabilityAgentic5.974.015 Sept 2026LMArena
Agent Arena task outcomeAgentic1.969.515 Sept 2026LMArena
ARC-AGI-3 (semi-private)Reasoning0.3%high effortARC Prize Foundation
ProgramBenchCoding0.0%21 Sept 2026Vals AI

43 benchmarks count, from 48 of 50 results. A grey row does not count. Too few models took that benchmark.

Sources

BenchLM benchmark aggregationUsed with attribution; per-benchmark results credited to their original authorsOpenRouter, collected directlyNo licence statedEpoch AI, collected directlyCC BY — free to use and redistribute with attributionVals AI, collected directlyNo licence stated. Read from the public leaderboard and credited to Vals AILiveBench, collected directlyNo licence stated for the leaderboard. The site repo serving these CSVs has no LICENSE; the harness repo carries upstream Apache-2.0 and MIT copies that cover code, not resultsARC Prize Foundation, collected directlyNo licence stated. Their terms ask for written permission before commercial useSierra Research, collected directlyMIT — results are in the licensed repositoryTerminal-Bench, collected directlyNo licence stated for the leaderboard. The harness repo is Apache-2.0LMArena, collected directlyCC BY 4.0 (lmarena-ai/leaderboard-dataset on Hugging Face)

Same level, lower price

Qwen3.8-Flash-Next66.4 · Freedots3-note Preview65.5 · FreeDeepSeek V4 Flash 042363.0 · $0.177

More from xAI

Grok 4.671.8Grok 4.770.1Grok 4.356.0Grok 4.20 (Reasoning)56.9Grok 451.6