Compare · ModelsLive · 2 picked · head to head
Grok 4.20 vs Claude Opus 4.5
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
Grok 4.20 wins on 9/14 benchmarks
Grok 4.20 wins 9 of 14 shared benchmarks. Leads in reasoning · knowledge · general.
Category leads
reasoning·Grok 4.20knowledge·Grok 4.20general·Grok 4.20math·Grok 4.20coding·Claude Opus 4.5
Hype vs Reality
Attention vs performance
Grok 4.20
#129 by perf·#14 by attention
Claude Opus 4.5
#179 by perf·#6 by attention
Best value
Grok 4.20
9.4x better value than Claude Opus 4.5
Grok 4.20
26.0 pts/$
$1.88/M
Claude Opus 4.5
2.8 pts/$
$15.00/M
Vendor risk
Who is behind the model
xAI
$250.0B·Tier 1
Anthropic
$965.0B·Tier 1
Head to head
14 benchmarks · 2 models
Grok 4.20Claude Opus 4.5
ARC-AGI
Grok 4.20 leads by +9.5
ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization.
Grok 4.20
89.5
Claude Opus 4.5
80.0
ARC-AGI-2
Grok 4.20 leads by +27.5
ARC-AGI-2 · the second iteration of the Abstraction and Reasoning Corpus, testing novel pattern recognition and abstract reasoning without prior training data.
Grok 4.20
65.1
Claude Opus 4.5
37.6
Chess Puzzles
Grok 4.20 leads by +12.6
Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities.
Grok 4.20
20.0
Claude Opus 4.5
7.4
Cl Bench
Grok 4.20 leads by +1.1
Grok 4.20
22.2
Claude Opus 4.5
21.1
Dtbench
Grok 4.20 leads by +0.4
Grok 4.20
83.5
Claude Opus 4.5
83.1
FrontierMath-Tier-4-v2-Private
Grok 4.20 leads by +12.2
Grok 4.20
17.1
Claude Opus 4.5
4.9
FrontierMath-Tiers-1-3-v2-Private
Grok 4.20 leads by +10.5
Grok 4.20
44.9
Claude Opus 4.5
34.4
GPQA diamond
Grok 4.20 leads by +4.4
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
Grok 4.20
85.8
Claude Opus 4.5
81.4
Lmca
Claude Opus 4.5 leads by +6.8
Grok 4.20
45.5
Claude Opus 4.5
52.3
OTIS Mock AIME 2024-2025
Grok 4.20 leads by +6.1
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
Grok 4.20
92.2
Claude Opus 4.5
86.1
Proofbench
Claude Opus 4.5 leads by +22.0
Grok 4.20
14.0
Claude Opus 4.5
36.0
SimpleQA Verified
Claude Opus 4.5 leads by +11.6
SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information.
Grok 4.20
30.2
Claude Opus 4.5
41.8
Terminal Bench
Claude Opus 4.5 leads by +5.9
Terminal-Bench 2.0 · evaluates AI agents on real terminal-based coding tasks · writing scripts, debugging, running tests, and managing projects entirely through command-line interaction. Tests both code quality and terminal fluency. Claude Opus 4.7 scores 69.4%, demonstrating significant agentic terminal competence.
Grok 4.20
57.3
Claude Opus 4.5
63.1
WeirdML
Claude Opus 4.5 leads by +11.5
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
Grok 4.20
52.3
Claude Opus 4.5
63.7
Full benchmark table
| Benchmark | Grok 4.20 | Claude Opus 4.5 |
|---|---|---|
ARC-AGI ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization. | 89.5 | 80.0 |
ARC-AGI-2 ARC-AGI-2 · the second iteration of the Abstraction and Reasoning Corpus, testing novel pattern recognition and abstract reasoning without prior training data. | 65.1 | 37.6 |
Chess Puzzles Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities. | 20.0 | 7.4 |
Cl Bench | 22.2 | 21.1 |
Dtbench | 83.5 | 83.1 |
FrontierMath-Tier-4-v2-Private | 17.1 | 4.9 |
FrontierMath-Tiers-1-3-v2-Private | 44.9 | 34.4 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 85.8 | 81.4 |
Lmca | 45.5 | 52.3 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 92.2 | 86.1 |
Proofbench | 14.0 | 36.0 |
SimpleQA Verified SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information. | 30.2 | 41.8 |
Terminal Bench Terminal-Bench 2.0 · evaluates AI agents on real terminal-based coding tasks · writing scripts, debugging, running tests, and managing projects entirely through command-line interaction. Tests both code quality and terminal fluency. Claude Opus 4.7 scores 69.4%, demonstrating significant agentic terminal competence. | 57.3 | 63.1 |
WeirdML WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns. | 52.3 | 63.7 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $1.25 | $2.50 | 2.0M tokens (~1,000 books) | $15.63 | |
| $5.00 | $25.00 | 200K tokens (~100 books) | $100.00 |
People also compared