Grok 4.3
xAI·Released Apr 17, 2026
Updated Oct 8, 11:55 PM ET
The score is on the board's scale (reference average 50, about 15 points per standard deviation), not a percentage.
By domain
Each domain is scored on the same scale as the overall score. Coverage is the share of the domain's kinds of evaluation with a result; it says how much evidence there is, not how good the model is.
Every result behind the score
The result as the evaluator publishes it, and the reading it gives on the board's scale. A hard evaluation can give a high reading for a modest result; results near 95% or 5% only say "at least" or "at most".
Coding
6 results · 27.0- ProgramBenchVals AI · high effortNear the floor0.0%Reading at most 65.7
- SciCodeArtificial Analysis · high effort48.3%Reading 45.0
- Terminal-Bench 4.0 (AA run)Artificial Analysis · high effortNear the floor0.0%Reading at most 34.5
- Vibe Code BenchVals AI · high effort19.4%Reading 29.3
- Code MigrationVals AI · high effort6.8%Reading 18.0
- LiveBench CodingLiveBench44.2%Reading 3.9
Research and reasoning
5 results · 39.7- Humanity's Last Exam (AA run)Artificial Analysis · high effort37.2%Reading 52.8
- Chess PuzzlesEpoch AI · high effort25.0%Reading 39.1
- FrontierMath Tier 4Epoch AI · high effort14.6%Reading 38.0
- FrontierMath Tiers 1–3Epoch AI · high effort42.8%Reading 36.7
- LiveBench ReasoningLiveBench77.6%Reading 33.1
Professional work
9 results · 21.1- τ²-Bench Telecom (AA run)Artificial Analysis · high effortNear the ceiling97.7%Reading at least 60.0
- Finance AgentVals AI · high effort37.7%Reading 29.3
- MedCodeVals AI38.1%Reading 27.2
- Legal Research BenchVals AI · high effort15.4%Reading 26.4
- Harvey Legal Agent BenchmarkVals AI · high effortNear the floor0.4%Reading at most 24.8
- τ-Bench Banking (AA run)Artificial Analysis · high effort12.4%Reading 14.9
- Tax Agent BenchVals AI · high effort6.2%Reading 6.8
- LiveBench Data AnalysisLiveBench55.8%Reading 6.8
- EMBVals AI · high effort18.4%Reading 4.9
Knowledge and accuracy
6 results · 46.4- BullshitBenchBullshitBench · xhigh effortCapped56.4%Reading 98.3
- IFBench (AA run)Artificial Analysis · high effort81.3%Reading 69.2
- LiveBench LanguageLiveBench73.6%Reading 41.4
- AA-LCRArtificial Analysis · high effort73.0%Reading 38.5
- LiveBench Instruction FollowingLiveBench62.8%Reading 34.1
- SimpleQA VerifiedEpoch AI · high effort33.2%Reading 30.0
Human preference
1 result · 36.1- Arena TextLMArena1,441.4Reading 38.0
Shown, not scored
Composite indices, saturated or older evaluations, evaluations still being checked, and vision, writing and multilingual results. They do not move the score.
- LiveCodeBenchVals AI · CodingReference only84.5%
- SkillsBenchVals AI · CodingWatching40.6%
- SWE-bench VerifiedVals AI · CodingReference only71.4%
- OTIS Mock AIMEEpoch AI · Research and reasoningReference only93.3%
- CorpFinVals AI · Professional workReference only68.5%
- LegalBenchVals AI · Professional workReference only84.5%
- MedScribeVals AI · Professional workWatching74.4%
- TaxEvalVals AI · Professional workReference only70.8%
- GPQA DiamondEpoch AI · Knowledge and accuracyReference only88.8%
- GPQA DiamondVals AI · Knowledge and accuracyReference only91.4%
- MMLU-ProVals AI · Knowledge and accuracyReference only85.8%
- AA Intelligence IndexArtificial Analysis · Composite indicesReference only24.9
- LiveBench averageLiveBench · Composite indicesReference only62.2%
- Arena VisionLMArena · VisionReference only1,242.9
- MMMU ProVals AI · VisionReference only83.1%
- Arena WebDevLMArena · Writing and designReference only1,356.9
No result yet
Scored evaluations this model has not taken. A missing result neither adds nor subtracts.
- Terminal-Bench 4.0 · Vals AI
- Vibe Code Bench 1–100 · Vals AI
- CyberBench Patch · Vals AI
- APEX-SWE · Mercor
- Mystery Game Puzzles · Epoch AI
- MysteryMechanism · Vals AI
- Terminal-Bench Science · Vals AI
- ProofBench · Vals AI
- APEX-Agents · Mercor