# SpaceXAI Grok 4.20

As of October 21, 2026, SpaceXAI Grok 4.20 ranks #67 of 213 models overall with 25.7 on the Artificial Analysis Intelligence Index. Its strongest category is reasoning, where it ranks #52 of 203. SpaceXAI Grok 4.20 costs $1.25 per million input tokens and $2.50 per million output tokens. It was released on March 31, 2026.

Page: https://llmbenchmark.ai/models/grok-4-20
Data from October 21, 2026. Every figure below names the board it was read from and the day.

- Maker: xAI
- Released: March 31, 2026
- Price per million input tokens: $1.25
- Price per million output tokens: $2.50
- Context window: 2M tokens
- Overall rank: 67 of 213, by the Artificial Analysis Intelligence Index
- Artificial Analysis Intelligence Index: 25.7

Benchmark results (54 of 190):

- GPQA Diamond: 88.6% (source: GPQA (Vals AI), read October 21, 2026)
- AIME: 96.5% (source: AIME (Vals AI), read October 21, 2026)
- MMLU Pro: 86.3% (source: MMLU Pro (Vals AI), read October 21, 2026)
- AA Intelligence Index: 25.7 (source: Artificial Analysis (via OpenRouter), read October 21, 2026)
- Design Arena: websites: 1231 Elo (source: Design Arena, read October 21, 2026)
- Design Arena: UI components: 1199 Elo (source: Design Arena, read October 21, 2026)
- Design Arena: game dev: 1192 Elo (source: Design Arena, read October 21, 2026)
- Design Arena: data viz: 1217 Elo (source: Design Arena, read October 21, 2026)
- Design Arena: 3D: 1194 Elo (source: Design Arena, read October 21, 2026)
- Design Arena: all code categories: 1222 Elo (source: Design Arena, read October 21, 2026)
- Design Arena: SVG: 1164 Elo (source: Design Arena, read October 21, 2026)
- Design Arena: ASCII art: 1175 Elo (source: Design Arena, read October 21, 2026)
- Design Arena: agents building web apps: 1143 Elo (source: Design Arena, read October 21, 2026)
- Design Arena: agents building full-stack apps: 1054 Elo (source: Design Arena, read October 21, 2026)
- Design Arena: agents building mobile apps: 1089 Elo (source: Design Arena, read October 21, 2026)
- Design Arena: agents building Android apps: 1051 Elo (source: Design Arena, read October 21, 2026)
- Design Arena: agents building Godot games: 1089 Elo (source: Design Arena, read October 21, 2026)
- Design Arena: HTML slides: 1180 Elo (source: Design Arena, read October 21, 2026)
- LiveCodeBench (Vals AI): 84.3% (source: Vals AI boards, read October 21, 2026)
- SWE-bench Verified (Vals AI): 72.2% (source: Vals AI boards, read October 21, 2026)
- Vibe Code Bench (Vals AI): 4.1% (source: Vals AI boards, read October 21, 2026)
- Code migration (Vals AI): 0.3% (source: Vals AI boards, read October 21, 2026)
- Terminal-Bench 2.1 (Vals AI): 44.2% (source: Vals AI boards, read October 21, 2026)
- Finance Agent (Vals AI): 28.5% (source: Vals AI boards, read October 21, 2026)
- Tax Agent Bench (Vals AI): 7.0% (source: Vals AI boards, read October 21, 2026)
- Legal Research Bench (Vals AI): 13.9% (source: Vals AI boards, read October 21, 2026)
- Excel Modeling Benchmark (Vals AI): 11.7% (source: Vals AI boards, read October 21, 2026)
- MMMU Pro (Vals AI): 83.5% (source: Vals AI boards, read October 21, 2026)
- SAGE (Vals AI): 38.2% (source: Vals AI boards, read October 21, 2026)
- LegalBench (Vals AI): 77.7% (source: Vals AI boards, read October 21, 2026)
- CaseLaw (Vals AI): 54.4% (source: Vals AI boards, read October 21, 2026)
- CorpFin (Vals AI): 63.7% (source: Vals AI boards, read October 21, 2026)
- TaxEval (Vals AI): 74.1% (source: Vals AI boards, read October 21, 2026)
- MortgageTax (Vals AI): 45.4% (source: Vals AI boards, read October 21, 2026)
- MedQA (Vals AI): 94.5% (source: Vals AI boards, read October 21, 2026)
- MedCode (Vals AI): 32.2% (source: Vals AI boards, read October 21, 2026)
- MedScribe (Vals AI): 63.4% (source: Vals AI boards, read October 21, 2026)
- Vals Multimodal Index: 39.1% (source: Vals AI boards, read October 21, 2026)
- GPQA Diamond (Artificial Analysis run): 91.1% (source: Artificial Analysis, read October 21, 2026)
- Humanity's Last Exam (Artificial Analysis run): 34.5% (source: Artificial Analysis, read October 21, 2026)
- Terminal-Bench Hard (Artificial Analysis run): 37.9% (source: Artificial Analysis, read October 21, 2026)
- IFBench: 81.2% (source: Artificial Analysis, read October 21, 2026)
- Long context reasoning (AA-LCR): 69.0% (source: Artificial Analysis, read October 21, 2026)
- τ²-Bench Telecom: 93.0% (source: Artificial Analysis, read October 21, 2026)
- Terminal-Bench: 57.3% (source: Epoch AI benchmarking hub, read October 21, 2026)
- WeirdML: 52.3% (source: Epoch AI benchmarking hub, read October 21, 2026)
- ARC-AGI: 89.5% (source: Epoch AI benchmarking hub, read October 21, 2026)
- ARC-AGI-2: 65.1% (source: Epoch AI benchmarking hub, read October 21, 2026)
- DTBench: 90.1% (source: Epoch AI benchmarking hub, read October 21, 2026)
- LMCA: 38.7% (source: Epoch AI benchmarking hub, read October 21, 2026)
- CL-bench: 22.2% (source: Epoch AI benchmarking hub, read October 21, 2026)
- CL-bench Life: 11.9% (source: Epoch AI benchmarking hub, read October 21, 2026)
- Vending-Bench 2: $4,663 (source: Epoch AI benchmarking hub, read October 21, 2026)
- Epoch Capabilities Index: 152.1 (source: Epoch AI benchmarking hub, read October 21, 2026)