LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

Gemini 2.5 Pro vs Grok 4.3

Updated October 21, 2026. 72 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.

Replace Gemini 2.5 Pro Replace Grok 4.3
Gemini 2.5 ProGoogle
Coding
37.0
Overall
16.1
Output $/M
$10.00
43
Grok 4.3xAI
Coding
22.5
Overall
24.9
Output $/M
$2.50
29
vs boards won of 72 shared

Decision reading

Pick Gemini 2.5 Pro for coding: it scores 37.0 on the Vibe Coding Index against 22.5 for Grok 4.3.

Pick Grok 4.3 when output tokens cost 4.0x less, or when a solved task costs $0.31 against $0.64.

Model AGemini 2.5 ProGoogle37.0/1002 independent benchmarks$0.64 per solved taskModel BGrok 4.3xAI22.5/1001 independent benchmark$0.31 per solved task
Shared results
72
Gemini 2.5 Pro only
15
Grok 4.3 only
32
Like-for-like categories
8 of 9

Shared results can support a head-to-head reading. Results only one model has describe coverage, not superiority.

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
Gemini 2.5 Pro 37.0 from 2 benchmarks, Grok 4.3 22.5 from 1 benchmark.
Widest category
Agentic work, 20.3 points apart on average over 3 shared benchmarks.
Widest single gap
τ²-Bench Telecom: Gemini 2.5 Pro 54.1% against Grok 4.3 97.7%, 43.6 points apart on a 0 to 100 scale.
Output price
Gemini 2.5 Pro $10.00 against Grok 4.3 $2.50 per million output tokens, 4.0x apart.
Context window
Gemini 2.5 Pro takes 1.05M tokens, Grok 4.3 1M.

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
CategoryGemini 2.5 ProBarsGrok 4.3Ties
Coding38
Intelligence01
Agentic work12
Reasoning53
Math23
Knowledge62
Web design15
General255
All shared boards4329
CodingLike-for-likeFour of these boards feed the Vibe Coding Index38
Coding benchmarks for Gemini 2.5 Pro and Grok 4.3
BenchmarkGemini 2.5 ProGrok 4.3Leader
SWE-bench Verified53.6%Not measuredOnly Gemini 2.5 Pro measured
LMArena WebDev12261358Grok 4.3
AA Coding Index33.342.2Grok 4.3
LiveCodeBench (Vals AI)Not measured84.5%Only Grok 4.3 measured
SWE-bench Verified (Vals AI)54.4%71.4%Grok 4.3
Vibe Code Bench (Vals AI)0.4%19.4%Grok 4.3
Code migration (Vals AI)Not measured6.8%Only Grok 4.3 measured
ProgramBench (Vals AI)Not measured0.0%Only Grok 4.3 measured
Terminal-Bench 2.1 (Vals AI)Not measured42.0%Only Grok 4.3 measured
SkillsBench (Vals AI)Not measured40.6%Only Grok 4.3 measured
LiveCodeBench (Artificial Analysis run)80.1%Not measuredOnly Gemini 2.5 Pro measured
SciCode46.3%48.3%Grok 4.3
Terminal-Bench Hard (Artificial Analysis run)26.5%37.9%Grok 4.3
Terminal-Bench32.6%Not measuredOnly Gemini 2.5 Pro measured
GSO-Bench3.9%Not measuredOnly Gemini 2.5 Pro measured
WeirdML54.0%49.9%Gemini 2.5 Pro
ALE-BenchNot measured944.2Only Grok 4.3 measured
LMArena Text: coding14521415Gemini 2.5 Pro
LMArena Text: software and IT14591416Gemini 2.5 Pro
LMArena WebDev: ReactNot measured1353Only Grok 4.3 measured
LMArena WebDev: HTML12351367Grok 4.3
LMArena WebDev: from an image12851375Grok 4.3
Neither model has a result on the other 16 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index01
Intelligence benchmarks for Gemini 2.5 Pro and Grok 4.3
BenchmarkGemini 2.5 ProGrok 4.3Leader
AA Intelligence Index16.124.9Grok 4.3
Epoch Capabilities Index145.1Not measuredOnly Gemini 2.5 Pro measured
Agentic workLike-for-likePublic boards, shown as published, kept out of the index12
Agentic work benchmarks for Gemini 2.5 Pro and Grok 4.3
BenchmarkGemini 2.5 ProGrok 4.3Leader
AA Agentic Index1.615.5Grok 4.3
Design Arena: agents building web appsNot measured1126Only Grok 4.3 measured
Design Arena: agents building full-stack appsNot measured1006Only Grok 4.3 measured
Design Arena: agents building mobile appsNot measured1059Only Grok 4.3 measured
Design Arena: agents building Android appsNot measured967Only Grok 4.3 measured
Design Arena: agents building Godot gamesNot measured914Only Grok 4.3 measured
Design Arena: agents building gamesNot measured1007Only Grok 4.3 measured
Design Arena: HTML slidesNot measured1008Only Grok 4.3 measured
Design Arena: PowerPoint slidesNot measured1012Only Grok 4.3 measured
Finance Agent (Vals AI)Not measured37.7%Only Grok 4.3 measured
Tax Agent Bench (Vals AI)Not measured6.2%Only Grok 4.3 measured
Legal Research Bench (Vals AI)Not measured15.4%Only Grok 4.3 measured
Excel Modeling Benchmark (Vals AI)Not measured18.4%Only Grok 4.3 measured
τ²-Bench Telecom54.1%97.7%Grok 4.3
GDPval23.3%Not measuredOnly Gemini 2.5 Pro measured
DeepResearch Bench41.5%Not measuredOnly Gemini 2.5 Pro measured
Vending-Bench 2$574$35Gemini 2.5 Pro
LMArena SearchNot measured1166Only Grok 4.3 measured
Neither model has a result on the other 21 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index53
Reasoning benchmarks for Gemini 2.5 Pro and Grok 4.3
BenchmarkGemini 2.5 ProGrok 4.3Leader
GPQA DiamondNot measured91.4%Only Grok 4.3 measured
MMMU Pro (Vals AI)Not measured83.1%Only Grok 4.3 measured
GPQA Diamond (Artificial Analysis run)84.4%90.1%Grok 4.3
Humanity's Last Exam (Artificial Analysis run)22.5%37.2%Grok 4.3
GPQA Diamond (Epoch AI run)85.3%88.8%Grok 4.3
ARC-AGI41.0%Not measuredOnly Gemini 2.5 Pro measured
ARC-AGI-24.9%Not measuredOnly Gemini 2.5 Pro measured
Chess puzzlesNot measured18.0%Only Grok 4.3 measured
LMArena Text: hard prompts14551396Gemini 2.5 Pro
LMArena Text: expert14521384Gemini 2.5 Pro
LMArena Text: hard prompts in English14531402Gemini 2.5 Pro
LMArena Vision: diagrams12741241Gemini 2.5 Pro
LMArena Vision: homework12811241Gemini 2.5 Pro
Neither model has a result on the other 8 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index23
Math benchmarks for Gemini 2.5 Pro and Grok 4.3
BenchmarkGemini 2.5 ProGrok 4.3Leader
AA Math Index87.7Not measuredOnly Gemini 2.5 Pro measured
AIME 2025 (Artificial Analysis run)87.7%Not measuredOnly Gemini 2.5 Pro measured
OTIS Mock AIME84.7%93.3%Grok 4.3
FrontierMath, tiers 1 to 324.6%42.8%Grok 4.3
FrontierMath, tier 40.0%14.6%Grok 4.3
ProofBenchNot measured11.0%Only Grok 4.3 measured
LMArena Text: math14501387Gemini 2.5 Pro
LMArena Text: mathematical work14541396Gemini 2.5 Pro
Neither model has a result on the other 5 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index62
Knowledge benchmarks for Gemini 2.5 Pro and Grok 4.3
BenchmarkGemini 2.5 ProGrok 4.3Leader
MMLU ProNot measured85.8%Only Grok 4.3 measured
LegalBench (Vals AI)Not measured84.5%Only Grok 4.3 measured
CaseLaw (Vals AI)63.9%79.3%Grok 4.3
CorpFin (Vals AI)60.8%68.5%Grok 4.3
TaxEval (Vals AI)Not measured70.8%Only Grok 4.3 measured
MortgageTax (Vals AI)68.9%48.3%Gemini 2.5 Pro
MedCode (Vals AI)50.6%38.1%Gemini 2.5 Pro
MMLU Pro (Artificial Analysis run)86.2%Not measuredOnly Gemini 2.5 Pro measured
SimpleQA VerifiedNot measured33.2%Only Grok 4.3 measured
LMArena Text: science14801399Gemini 2.5 Pro
LMArena Text: medicine and healthcare14761384Gemini 2.5 Pro
LMArena Text: legal and government14801386Gemini 2.5 Pro
LMArena Text: business and finance14391381Gemini 2.5 Pro
Neither model has a result on the other 4 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index15
Web design benchmarks for Gemini 2.5 Pro and Grok 4.3
BenchmarkGemini 2.5 ProGrok 4.3Leader
Design Arena: websites11671196Grok 4.3
Design Arena: UI components11391192Grok 4.3
Design Arena: game dev11121174Grok 4.3
Design Arena: data viz12221180Gemini 2.5 Pro
Design Arena: 3D10801125Grok 4.3
Design Arena: all code categories11521187Grok 4.3
Design Arena: SVGNot measured1081Only Grok 4.3 measured
Design Arena: ASCII artNot measured1125Only Grok 4.3 measured
Neither model has a result on the other 1 board here
GeneralLike-for-likePublic boards, shown as published, kept out of the index255
General benchmarks for Gemini 2.5 Pro and Grok 4.3
BenchmarkGemini 2.5 ProGrok 4.3Leader
Public Benefits Bench (Vals AI)Not measured51.7%Only Grok 4.3 measured
SAGE (Vals AI)41.9%19.7%Gemini 2.5 Pro
MedScribe (Vals AI)73.5%74.4%Grok 4.3
Vals Multimodal IndexNot measured43.3%Only Grok 4.3 measured
IFBench48.7%83.3%Grok 4.3
Long context reasoning (AA-LCR)69.0%75.0%Grok 4.3
DTBench82.4%90.7%Grok 4.3
LMCA34.8%38.3%Grok 4.3
Creative writing (Lech Mazur)83.8%Not measuredOnly Gemini 2.5 Pro measured
LMArena Text14581397Gemini 2.5 Pro
LMArena Text: instruction following14371366Gemini 2.5 Pro
LMArena Text: creative writing14541380Gemini 2.5 Pro
LMArena Vision12631229Gemini 2.5 Pro
LMArena Text: English14581404Gemini 2.5 Pro
LMArena Text: non-English14521385Gemini 2.5 Pro
LMArena Text: Chinese15071422Gemini 2.5 Pro
LMArena Text: German14871395Gemini 2.5 Pro
LMArena Text: Japanese14611379Gemini 2.5 Pro
LMArena Text: Korean14341354Gemini 2.5 Pro
LMArena Text: French14731411Gemini 2.5 Pro
LMArena Text: Spanish14741399Gemini 2.5 Pro
LMArena Text: Russian14611399Gemini 2.5 Pro
LMArena Text: Polish14791398Gemini 2.5 Pro
LMArena Text: multi-turn14531406Gemini 2.5 Pro
LMArena Text: longer queries14491393Gemini 2.5 Pro
LMArena Text: writing and language14541385Gemini 2.5 Pro
LMArena Text: entertainment and media14411370Gemini 2.5 Pro
LMArena Vision: English12611235Gemini 2.5 Pro
LMArena Vision: Chinese13001257Gemini 2.5 Pro
LMArena Vision: reading text12701235Gemini 2.5 Pro
LMArena Vision: humor12621222Gemini 2.5 Pro
LMArena Vision: captioning1304Not measuredOnly Gemini 2.5 Pro measured
LMArena Vision: recognizing things12831217Gemini 2.5 Pro
LMArena Vision: creative writing12651237Gemini 2.5 Pro
LMArena Document1421Not measuredOnly Gemini 2.5 Pro measured
Neither model has a result on the other 11 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Gemini 2.5 Pro and Grok 4.3
BenchmarkGemini 2.5 ProGrok 4.3Leader
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

Gemini 2.5 Pro$0.0063Fits in one request
Grok 4.3$0.0025Fits in one request

Grok 4.3 costs less

Repository review

50K input + 3K output tokens

Gemini 2.5 Pro$0.092Fits in one request
Grok 4.3$0.070Fits in one request

Grok 4.3 costs less

Agent loop

200K input + 10K output tokens

Gemini 2.5 Pro$0.350Fits in one request
Grok 4.3$0.275Fits in one request

Grok 4.3 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 Gemini 2.5 Pro and Grok 4.3
SpecificationGoogle Gemini 2.5 ProSpaceXAI Grok 4.3
MakerGooglexAI
WeightsNo public weights listedNo public weights listed
ReleasedJune 17, 2025April 30, 2026
Context window1.05M1M
Input $/Mtok$1.25$1.25
Output $/Mtok$10.00$2.50
Coding index37.022.5
Independent measurements behind it21
Cost per solved task$0.64$0.31
SourceOpenRouter listingOpenRouter listing

Questions

Is Gemini 2.5 Pro better than Grok 4.3 for coding?

Pick Gemini 2.5 Pro for coding: it scores 37.0 on the Vibe Coding Index against 22.5 for Grok 4.3. Pick Grok 4.3 when output tokens cost 4.0x less, or when a solved task costs $0.31 against $0.64.

Which is cheaper, Gemini 2.5 Pro or Grok 4.3?

Gemini 2.5 Pro costs $1.25 per million input tokens and $10.00 per million output tokens; Grok 4.3 costs $1.25 and $2.50. Per solved SWE-bench Verified task, Gemini 2.5 Pro costs $0.64 and Grok 4.3 $0.31.

Which has the bigger context window, Gemini 2.5 Pro or Grok 4.3?

Gemini 2.5 Pro, with 1.05M tokens against 1M.

How many benchmarks is the Gemini 2.5 Pro vs Grok 4.3 comparison based on?

72 benchmarks have a result for both models as of October 21, 2026. Gemini 2.5 Pro's coding index rests on 2 independent benchmarks and Grok 4.3's on 1. Where only one model has a result, the page says so instead of naming a winner.

Compare next