Enter opens the highlighted result. Esc closes.

Claude Opus 4.5 vs Grok 4.20

Claude Opus 4.5 leads Grok 4.20 on coding, 44.0 to 33.3 on the Vibe Coding Index, and Grok 4.20 costs 10.0x less per output token ($2.50 vs $25.00 per million). Updated October 21, 2026; 45 benchmarks have a result for both.

Replace Claude Opus 4.5 Replace Grok 4.20
Claude Opus 4.5Anthropic
Coding
44.0
Overall
55.7
Output $/M
$25.00
32
Grok 4.20xAI
Coding
33.3
Overall
41.0
Output $/M
$2.50
13
vs boards won of 45 shared

Claude Opus 4.5 or Grok 4.20: which to pick

Pick Claude Opus 4.5 for coding: it scores 44.0 on the Vibe Coding Index against 33.3 for Grok 4.20.

Pick Grok 4.20 for output tokens 10.0x cheaper ($2.50 vs $25.00 per million) and a lower cost per solved task ($0.72 against $1.37).

Model AClaude Opus 4.5Anthropic44.0/1003 independent benchmarks$1.37 per solved taskModel BGrok 4.20xAI33.3/1002 independent benchmarks$0.72 per solved task
Shared results
45
Claude Opus 4.5 only
63
Grok 4.20 only
10
Like-for-like categories
8 of 9

Only benchmarks that tested both models count toward a pick; a result only one model has is shown but not compared.

Which one for your work

Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.

At a glance

Coding index
Claude Opus 4.5 44.0 from 3 benchmarks, Grok 4.20 33.3 from 2 benchmarks.
Widest category
General, 10.6 points apart on average over 7 shared benchmarks.
Widest single gap
ARC-AGI-2: Claude Opus 4.5 37.6% against Grok 4.20 65.1%, 27.5 points apart on a 0 to 100 scale.
Output price
Claude Opus 4.5 $25.00 against Grok 4.20 $2.50 per million output tokens, 10.0x apart.
Context window
Claude Opus 4.5 takes 200K tokens, Grok 4.20 2M.

Benchmark by benchmark

First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.

Boards won, where both have a result
CategoryClaude Opus 4.5BarsGrok 4.20Ties
Coding51
Intelligence11
Agentic work51
Reasoning16
Math01
Knowledge80
Web design80
General43
All shared boards3213
CodingLike-for-likeSix of these boards feed the Vibe Coding Index51
Coding benchmarks for Claude Opus 4.5 and Grok 4.20
BenchmarkClaude Opus 4.5Grok 4.20Leader
SWE-bench Verified79.2%Not measuredOnly Claude Opus 4.5 measured
SWE-bench (full test set)52.6%Not measuredOnly Claude Opus 4.5 measured
SWE-bench Multilingual70.7%Not measuredOnly Claude Opus 4.5 measured
LMArena WebDev1494Not measuredOnly Claude Opus 4.5 measured
LiveCodeBench (Vals AI)83.7%84.3%Grok 4.20
SWE-bench Verified (Vals AI)76.4%72.2%Claude Opus 4.5
Vibe Code Bench (Vals AI)20.6%4.1%Claude Opus 4.5
Code migration (Vals AI)Not measured0.3%Only Grok 4.20 measured
Terminal-Bench 2.1 (Vals AI)Not measured44.2%Only Grok 4.20 measured
LiveCodeBench (Artificial Analysis run)87.1%Not measuredOnly Claude Opus 4.5 measured
Terminal-Bench Hard (Artificial Analysis run)47.0%37.9%Claude Opus 4.5
Terminal-Bench63.1%57.3%Claude Opus 4.5
GSO-Bench26.5%Not measuredOnly Claude Opus 4.5 measured
WeirdML63.7%52.3%Claude Opus 4.5
Cybench82.0%Not measuredOnly Claude Opus 4.5 measured
LMArena Text: coding1503Not measuredOnly Claude Opus 4.5 measured
LMArena Text: software and IT1483Not measuredOnly Claude Opus 4.5 measured
LMArena WebDev: React1478Not measuredOnly Claude Opus 4.5 measured
LMArena WebDev: HTML1510Not measuredOnly Claude Opus 4.5 measured
Neither model has a result on the other 20 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index11
Intelligence benchmarks for Claude Opus 4.5 and Grok 4.20
BenchmarkClaude Opus 4.5Grok 4.20Leader
AA Intelligence Index29.125.7Claude Opus 4.5
Epoch Capabilities Index150.1152.1Grok 4.20
Agentic workLike-for-likePublic boards, shown as published, kept out of the index51
Agentic work benchmarks for Claude Opus 4.5 and Grok 4.20
BenchmarkClaude Opus 4.5Grok 4.20Leader
Design Arena: agents building web apps11591144Claude Opus 4.5
Design Arena: agents building full-stack apps11561054Claude Opus 4.5
Design Arena: agents building mobile apps11891090Claude Opus 4.5
Design Arena: agents building Android apps11751052Claude Opus 4.5
Design Arena: agents building Godot gamesNot measured1089Only Grok 4.20 measured
Design Arena: HTML slidesNot measured1180Only Grok 4.20 measured
Finance Agent (Vals AI)Not measured28.5%Only Grok 4.20 measured
Tax Agent Bench (Vals AI)Not measured7.0%Only Grok 4.20 measured
Legal Research Bench (Vals AI)Not measured13.9%Only Grok 4.20 measured
Excel Modeling Benchmark (Vals AI)Not measured11.7%Only Grok 4.20 measured
τ²-Bench Telecom89.5%93.0%Grok 4.20
METR time horizons75.0%Not measuredOnly Claude Opus 4.5 measured
Remote Labor Index3.8%Not measuredOnly Claude Opus 4.5 measured
GDPval45.5%Not measuredOnly Claude Opus 4.5 measured
BALROG43.5%Not measuredOnly Claude Opus 4.5 measured
DeepResearch Bench54.8%Not measuredOnly Claude Opus 4.5 measured
OSWorld66.3%Not measuredOnly Claude Opus 4.5 measured
Vending-Bench 2$4,967$4,663Claude Opus 4.5
Berkeley Function Calling Leaderboard (BFCL)77.5%Not measuredOnly Claude Opus 4.5 measured
Neither model has a result on the other 22 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index16
Reasoning benchmarks for Claude Opus 4.5 and Grok 4.20
BenchmarkClaude Opus 4.5Grok 4.20Leader
GPQA Diamond85.9%88.6%Grok 4.20
MMMU Pro (Vals AI)83.0%83.5%Grok 4.20
GPQA Diamond (Artificial Analysis run)86.6%91.1%Grok 4.20
Humanity's Last Exam (Artificial Analysis run)30.1%34.5%Grok 4.20
GPQA Diamond (Epoch AI run)86.0%Not measuredOnly Claude Opus 4.5 measured
ARC-AGI80.0%89.5%Grok 4.20
ARC-AGI-237.6%65.1%Grok 4.20
SimpleBench62.0%Not measuredOnly Claude Opus 4.5 measured
Chess puzzles14.0%Not measuredOnly Claude Opus 4.5 measured
Mystery game puzzles22.0%Not measuredOnly Claude Opus 4.5 measured
EBR-bench14.3%Not measuredOnly Claude Opus 4.5 measured
VPCT40.0%Not measuredOnly Claude Opus 4.5 measured
LMArena Text: hard prompts1476Not measuredOnly Claude Opus 4.5 measured
LMArena Text: expert1487Not measuredOnly Claude Opus 4.5 measured
LMArena Text: hard prompts in English1482Not measuredOnly Claude Opus 4.5 measured
Kagi LLM Benchmark80.2%75.0%Claude Opus 4.5
Neither model has a result on the other 6 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index01
Math benchmarks for Claude Opus 4.5 and Grok 4.20
BenchmarkClaude Opus 4.5Grok 4.20Leader
AIME95.4%96.5%Grok 4.20
MGSM (Vals AI)95.2%Not measuredOnly Claude Opus 4.5 measured
AA Math Index91.3Not measuredOnly Claude Opus 4.5 measured
AIME 2025 (Artificial Analysis run)91.3%Not measuredOnly Claude Opus 4.5 measured
OTIS Mock AIME86.1%Not measuredOnly Claude Opus 4.5 measured
FrontierMath, tiers 1 to 334.4%Not measuredOnly Claude Opus 4.5 measured
FrontierMath, tier 44.9%Not measuredOnly Claude Opus 4.5 measured
ProofBench36.0%Not measuredOnly Claude Opus 4.5 measured
LMArena Text: math1463Not measuredOnly Claude Opus 4.5 measured
LMArena Text: mathematical work1468Not measuredOnly Claude Opus 4.5 measured
Neither model has a result on the other 3 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index80
Knowledge benchmarks for Claude Opus 4.5 and Grok 4.20
BenchmarkClaude Opus 4.5Grok 4.20Leader
MMLU Pro87.3%86.3%Claude Opus 4.5
LegalBench (Vals AI)84.6%77.7%Claude Opus 4.5
CaseLaw (Vals AI)62.6%54.4%Claude Opus 4.5
CorpFin (Vals AI)65.1%63.7%Claude Opus 4.5
TaxEval (Vals AI)74.9%74.1%Claude Opus 4.5
MortgageTax (Vals AI)68.7%45.4%Claude Opus 4.5
MedQA (Vals AI)95.9%94.5%Claude Opus 4.5
MedCode (Vals AI)49.2%32.2%Claude Opus 4.5
MMLU Pro (Artificial Analysis run)89.5%Not measuredOnly Claude Opus 4.5 measured
SimpleQA Verified45.7%Not measuredOnly Claude Opus 4.5 measured
GeoBench75.0%Not measuredOnly Claude Opus 4.5 measured
LMArena Text: science1454Not measuredOnly Claude Opus 4.5 measured
LMArena Text: medicine and healthcare1444Not measuredOnly Claude Opus 4.5 measured
LMArena Text: legal and government1459Not measuredOnly Claude Opus 4.5 measured
LMArena Text: business and finance1448Not measuredOnly Claude Opus 4.5 measured
FACTS Benchmark Suite43.9%Not measuredOnly Claude Opus 4.5 measured
Neither model has a result on the other 2 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index80
Web design benchmarks for Claude Opus 4.5 and Grok 4.20
BenchmarkClaude Opus 4.5Grok 4.20Leader
Design Arena: websites12471229Claude Opus 4.5
Design Arena: UI components12381199Claude Opus 4.5
Design Arena: game dev12291192Claude Opus 4.5
Design Arena: data viz12431217Claude Opus 4.5
Design Arena: 3D12141193Claude Opus 4.5
Design Arena: all code categories12421221Claude Opus 4.5
Design Arena: SVG11811164Claude Opus 4.5
Design Arena: ASCII art11901174Claude Opus 4.5
Neither model has a result on the other 1 board here
GeneralLike-for-likePublic boards, shown as published, kept out of the index43
General benchmarks for Claude Opus 4.5 and Grok 4.20
BenchmarkClaude Opus 4.5Grok 4.20Leader
SAGE (Vals AI)52.1%38.2%Claude Opus 4.5
MedScribe (Vals AI)85.3%63.4%Claude Opus 4.5
Vals Multimodal IndexNot measured39.1%Only Grok 4.20 measured
IFBench58.0%81.2%Grok 4.20
Long context reasoning (AA-LCR)77.3%69.0%Claude Opus 4.5
DTBench89.9%90.1%Grok 4.20
LMCA44.5%38.7%Claude Opus 4.5
CL-bench21.1%22.2%Grok 4.20
Furniture assembly28.3%Not measuredOnly Claude Opus 4.5 measured
CL-bench LifeNot measured11.9%Only Grok 4.20 measured
LMArena Text1451Not measuredOnly Claude Opus 4.5 measured
LMArena Text: instruction following1478Not measuredOnly Claude Opus 4.5 measured
LMArena Text: creative writing1445Not measuredOnly Claude Opus 4.5 measured
LMArena Text: English1461Not measuredOnly Claude Opus 4.5 measured
LMArena Text: non-English1438Not measuredOnly Claude Opus 4.5 measured
LMArena Text: Chinese1470Not measuredOnly Claude Opus 4.5 measured
LMArena Text: German1449Not measuredOnly Claude Opus 4.5 measured
LMArena Text: Japanese1417Not measuredOnly Claude Opus 4.5 measured
LMArena Text: Korean1424Not measuredOnly Claude Opus 4.5 measured
LMArena Text: French1471Not measuredOnly Claude Opus 4.5 measured
LMArena Text: Spanish1457Not measuredOnly Claude Opus 4.5 measured
LMArena Text: Russian1447Not measuredOnly Claude Opus 4.5 measured
LMArena Text: Polish1439Not measuredOnly Claude Opus 4.5 measured
LMArena Text: multi-turn1466Not measuredOnly Claude Opus 4.5 measured
LMArena Text: longer queries1479Not measuredOnly Claude Opus 4.5 measured
LMArena Text: writing and language1452Not measuredOnly Claude Opus 4.5 measured
LMArena Text: entertainment and media1434Not measuredOnly Claude Opus 4.5 measured
LMArena Document1462Not measuredOnly Claude Opus 4.5 measured
Neither model has a result on the other 18 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Claude Opus 4.5 and Grok 4.20
BenchmarkClaude Opus 4.5Grok 4.20Leader
Neither model has a result on 5 boards here

What each workload costs

Three fixed token mixes at each model's listed price, so per-token rates become a bill.

Chat turn

1K input + 500 output tokens

Claude Opus 4.5$0.018Fits in one request
Grok 4.20$0.0025Fits in one request

Grok 4.20 costs less

Repository review

50K input + 3K output tokens

Claude Opus 4.5$0.325Fits in one request
Grok 4.20$0.070Fits in one request

Grok 4.20 costs less

Agent loop

200K input + 10K output tokens

Claude Opus 4.5$1.25Exceeds the context window
Grok 4.20$0.275Fits in one request

Grok 4.20 costs less

Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.

Specification differences

The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.

Specifications of Claude Opus 4.5 and Grok 4.20
SpecificationAnthropic Claude Opus 4.5SpaceXAI Grok 4.20
MakerAnthropicxAI
WeightsNo public weights listedNo public weights listed
ReleasedNovember 24, 2025March 31, 2026
Context window200K2M
Input $/Mtok$5.00$1.25
Output $/Mtok$25.00$2.50
Coding index44.033.3
Independent measurements behind it32
Cost per solved task$1.37$0.72
SourceOpenRouter listingOpenRouter listing

Questions

Is Claude Opus 4.5 better than Grok 4.20 for coding?

Pick Claude Opus 4.5 for coding: it scores 44.0 on the Vibe Coding Index against 33.3 for Grok 4.20. Pick Grok 4.20 for output tokens 10.0x cheaper ($2.50 vs $25.00 per million) and a lower cost per solved task ($0.72 against $1.37).

Which is cheaper, Claude Opus 4.5 or Grok 4.20?

Claude Opus 4.5 costs $5.00 per million input tokens and $25.00 per million output tokens; Grok 4.20 costs $1.25 and $2.50. Per solved SWE-bench Verified task, Claude Opus 4.5 costs $1.37 and Grok 4.20 $0.72.

Which has the bigger context window, Claude Opus 4.5 or Grok 4.20?

Grok 4.20, with 2M tokens against 200K.

How many benchmarks is the Claude Opus 4.5 vs Grok 4.20 comparison based on?

45 benchmarks have a result for both models as of October 21, 2026. Claude Opus 4.5's coding index rests on 3 independent benchmarks and Grok 4.20's on 2. Where only one model has a result, the page says so instead of naming a winner.

Compare next

Embed this comparison on your site

Free under CC BY 4.0, while the source link stays visible. Preview the frame · Terms and other embeds

<iframe src="https://llmbenchmark.ai/embed/compare/claude-opus-4-5-vs-grok-4-20" title="Claude Opus 4.5 vs Grok 4.20, LLMBenchmark.ai" width="100%" height="360" style="border:0;max-width:100%" loading="lazy"></iframe>
<p style="font:13px/1.4 sans-serif;margin:6px 0 0">Source: <a href="https://llmbenchmark.ai/compare/claude-opus-4-5-vs-grok-4-20">LLMBenchmark.ai, Claude Opus 4.5 vs Grok 4.20</a></p>