xAI
Best on the Capabilities Index Grok 4.6 157.8
How does xAI stack up?
Every model on Epoch AI's Capabilities Index by release date, with xAI's in color.
Source: Epoch Capabilities Index, Epoch AI (CC BY 4.0)
xAI's best models
xAI Aug 2026
Grok 4.6
Closed weights Flagship
157.8
#28 of 256
Epoch Capabilities Index
Self-Reported Index 160.7 #26 of 404
- DeepSWE 67.5% #9 /28
- Terminal-Bench 4.0 20.3% #25 /40
- ZeroBench 17% #26 /76
xAI Sep 2026
Grok 4.7
Closed weights Flagship
156.2
#38 of 256
Epoch Capabilities Index
Self-Reported Index 157.3 #49 of 404
- Terminal-Bench 4.0 37.6% #16 /40
- GPQA Diamond 92.7% #27 /229
- HealthBench Professional 56.7% #21 /26
xAI Jul 2026
Grok 4.5
Closed weights Flagship
155.0
#46 of 256
Epoch Capabilities Index
Self-Reported Index 152.2 #81 of 404
- DeepSWE 53.8% #19 /28
- Agents' Last Exam 27% #12 /31
- Terminal-Bench 4.0 12.4% #32 /40
xAI Feb 2026
Grok 4.20
Closed weights
152.3
#60 of 256
Epoch Capabilities Index
- Terminal-Bench 2.0 57.3% #14 /43
- ZeroBench 10% #39 /76
- GPQA Diamond 89.3% #55 /229
xAI Apr 2026
Grok 4.3 Beta
Closed weights
149.9
#71 of 256
Epoch Capabilities Index
- GPQA Diamond 88.8% #57 /229
- Chess Puzzles 18% #75 /93
- OTIS Mock AIME 2024-2025 93.3% #59 /212
xAI Jul 2025
Grok 4
Closed weights
146.9
#85 of 256
Epoch Capabilities Index
Self-Reported Index 144.3 #137 of 404
- Terminal-Bench 2.0 27.2% #34 /43
- Toolathlon 27.5% #34 /50
- ZeroBench 6.00% #49 /76
What benchmarks does xAI mention most?
Here are the 25 benchmarks xAI mentions most in its launch posts.
MMLU HumanEval MATH GSM8K GPQA Diamond MMLU-Pro MMMU LMArena (Elo) AIME 2025 LiveCodeBench SimpleQA AIME 2024 LOFT EgoSchema Humanity's Last Exam HMMT 2025 USAMO 2025 Research-Eval X Browse (xAI) τ-Voice DeepSWE Big Bench Audio Speech-to-Speech Index Full-Duplex-Bench Harvey LAB (Legal Agent Benchmark)
Source: Labs’ launch posts, read by Benchmark Almanac
General refusals (xAI) MASK Virology Capabilities Test (VCT) WMDP Refusals + User Jailbreak (xAI) Refusals + System Jailbreak (xAI) AgentHarm AgentDojo Cybench SycophancyEval (answer sycophancy) MakeMeSay BioLP-bench Soft Bias (Internal) xAI sycophancy eval ProtocolQA Open-Ended (xAI) DeepSWE SWE-Marathon EEBench CyberGym HackerBench LAB-Bench BixBench Standard jailbreaks (xAI) Child safety (xAI) Autointent-Bio Autointent-Chem FORTRESS-RN Self-harm refusals (xAI)
Source: Labs’ system cards and technical reports, read by Benchmark Almanac
MMLU HumanEval MATH GPQA Diamond MMLU-Pro MMMU LMArena (Elo) AIME 2025 LiveCodeBench SimpleQA General refusals (xAI) MASK Virology Capabilities Test (VCT) WMDP Refusals + User Jailbreak (xAI) Refusals + System Jailbreak (xAI) AgentHarm AgentDojo Cybench SycophancyEval (answer sycophancy) MakeMeSay BioLP-bench xAI sycophancy eval ProtocolQA Open-Ended (xAI) DeepSWE SWE-Marathon EEBench Harvey LAB (Legal Agent Benchmark)
Source: Labs’ launch posts, system cards and technical reports, read by Benchmark Almanac
Model cards
| Date | Model | Document | Benchmarks |
|---|---|---|---|
| Oct 14, 2026 | Grok Voice 3.0 | Introducing Grok Voice 3.0 (xAI news) | 6 |
| Sep 21, 2026 | Grok 4.7 | Introducing Grok 4.7 | 8 |
| Aug 12, 2026 | Grok 4.6 | Introducing Grok 4.6 | 10 |
| Jul 29, 2026 | Grok Voice 2.0 | Introducing Grok Voice 2.0 (xAI news) | 4 |
| Jul 16, 2026 | Grok 4.5 | Introducing Grok 4.5 | 5 |
| Apr 23, 2026 | Grok Voice Think Fast 1.0 | Grok Voice Think Fast 1.0 (xAI news) | 1 |
| Mar 9, 2026 | Grok 4.20 | Grok 4.20 model documentation | 18 |
| Nov 19, 2025 | Grok 4.1 Fast | Grok 4.1 Fast and Agent Tools API (xAI news) | 5 |
| Nov 17, 2025 | Grok 4.1 | Grok 4.1 | 4 |
| Sep 19, 2025 | Grok 4 Fast | Grok 4 Fast (xAI news) | 12 |
| Aug 28, 2025 | grok-code-fast-1 | Grok Code Fast 1 (xAI news) | 1 |
| Jul 9, 2025 | Grok 4 | Grok 4 (xAI news) | 8 |
| Jul 9, 2025 | Grok 4 Heavy | Grok 4 | 6 |
| Feb 19, 2025 | Grok-3 mini | Grok 3 Beta: The Age of Reasoning Agents (xAI news) | 9 |
| Feb 19, 2025 | Grok 3 | Grok 3 Beta — The Age of Reasoning Agents | 10 |
| Aug 13, 2024 | Grok-2 Aug 2024 | Grok-2 Beta Release | 9 |
| Mar 28, 2024 | Grok-1.5 | Announcing Grok-1.5 | 4 |
| Nov 3, 2023 | Grok-1 | Announcing Grok | 5 |
Every xAI model
- Grok Voice 3.0 Oct 2026
- Grok 4.7 Sep 2026
- Grok 4.6 Aug 2026
- Grok Voice 2.0 Jul 2026
- Grok 4.5 Jul 2026
- Grok Build 0.1 May 2026
- Grok 4.3 Apr 2026
- Grok Voice Think Fast 1.0 Apr 2026
- Grok 4.3 Beta Apr 2026
- Grok TTS Apr 2026
- Grok 4.20 Multi-agent Mar 2026
- Grok 4.20 Feb 2026
- Grok 4.1 Fast Nov 2025
- Grok 4.1 Nov 2025
- Grok 4 Fast Sep 2025
- Grok 4 Heavy Jul 2025
- Grok 4 Jul 2025
- Grok 3 Apr 2025
- Grok-3 mini Apr 2025
- Grok-2 Dec 2024 Dec 2024
- Grok 2 Vision (1212) Dec 2024
- Grok-2 Aug 2024 Aug 2024
- Grok-2 mini Aug 2024
- Grok-1.5 Mar 2024
- Grok-1 Nov 2023
- Grok Imagine Image 2.0
- Grok Imagine Video
- Grok Imagine Video 1.5