LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

Grok 4.5 vs MiMo-V2.5

Updated October 21, 2026. 76 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.

Replace Grok 4.5 Replace MiMo-V2.5
Grok 4.5xAI
Coding
75.5
Overall
38.8
Output $/M
$6.00
72
MiMo-V2.5Xiaomi
Coding
34.0
Overall
21.8
Output $/M
$0.238
4
vs boards won of 76 shared

Decision reading

Pick Grok 4.5 for coding: it scores 75.5 on the Vibe Coding Index against 34.0 for MiMo-V2.5.

Pick MiMo-V2.5 when output tokens cost 25.2x less, or when a solved task costs $0.01 against $0.62.

Model AGrok 4.5xAI75.5/1002 independent benchmarks$0.62 per solved taskModel BMiMo-V2.5Xiaomi34.0/1001 independent benchmark$0.01 per solved task
Shared results
76
Grok 4.5 only
44
MiMo-V2.5 only
4
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
Grok 4.5 75.5 from 2 benchmarks, MiMo-V2.5 34.0 from 1 benchmark.
Widest category
Agentic work, 17.0 points apart on average over 5 shared benchmarks.
Widest single gap
Legal Research Bench (Vals AI): Grok 4.5 38.0% against MiMo-V2.5 9.1%, 28.8 points apart on a 0 to 100 scale.
Output price
Grok 4.5 $6.00 against MiMo-V2.5 $0.238 per million output tokens, 25.2x apart.
Context window
Grok 4.5 takes 500K tokens, MiMo-V2.5 1.05M.

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
CategoryGrok 4.5BarsMiMo-V2.5Ties
Coding130
Intelligence10
Agentic work41
Reasoning81
Math30
Knowledge91
Web design80
General261
All shared boards724
CodingLike-for-likeFour of these boards feed the Vibe Coding Index130
Coding benchmarks for Grok 4.5 and MiMo-V2.5
BenchmarkGrok 4.5MiMo-V2.5Leader
LiveBench coding94.3%Not measuredOnly Grok 4.5 measured
LMArena WebDev15531438Grok 4.5
AA Coding Index72.456.8Grok 4.5
LiveCodeBench (Vals AI)87.3%81.5%Grok 4.5
SWE-bench Verified (Vals AI)86.6%71.0%Grok 4.5
Terminal-Bench 4 (Vals AI)8.6%Not measuredOnly Grok 4.5 measured
Vibe Code Bench (Vals AI)69.0%42.2%Grok 4.5
Code migration (Vals AI)36.6%14.2%Grok 4.5
ProgramBench (Vals AI)0.0%Not measuredOnly Grok 4.5 measured
IOI (Vals AI)40.6%Not measuredOnly Grok 4.5 measured
Terminal-Bench 2.1 (Vals AI)67.8%60.7%Grok 4.5
SkillsBench (Vals AI)66.0%Not measuredOnly Grok 4.5 measured
CyberBench (Vals AI)71.0%Not measuredOnly Grok 4.5 measured
SciCode55.0%43.9%Grok 4.5
Terminal-Bench Hard (Artificial Analysis run)Not measured41.7%Only MiMo-V2.5 measured
FrontierCode42.4%Not measuredOnly Grok 4.5 measured
WeirdML46.4%Not measuredOnly Grok 4.5 measured
ALE-Bench1309.3514.0Grok 4.5
LMArena Text: coding14741469Grok 4.5
LMArena Text: software and IT14711463Grok 4.5
LMArena WebDev: React15561425Grok 4.5
LMArena WebDev: HTML15671429Grok 4.5
LMArena WebDev: from an image1562Not measuredOnly Grok 4.5 measured
Neither model has a result on the other 15 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index10
Intelligence benchmarks for Grok 4.5 and MiMo-V2.5
BenchmarkGrok 4.5MiMo-V2.5Leader
AA Intelligence Index38.821.8Grok 4.5
Epoch Capabilities Index154.1Not measuredOnly Grok 4.5 measured
Agentic workLike-for-likePublic boards, shown as published, kept out of the index41
Agentic work benchmarks for Grok 4.5 and MiMo-V2.5
BenchmarkGrok 4.5MiMo-V2.5Leader
AA Agentic Index41.215.8Grok 4.5
Design Arena: agents building web apps1217Not measuredOnly Grok 4.5 measured
Design Arena: agents building full-stack apps1249Not measuredOnly Grok 4.5 measured
Design Arena: agents building mobile apps1221Not measuredOnly Grok 4.5 measured
Design Arena: agents building Android apps1242Not measuredOnly Grok 4.5 measured
Design Arena: agents building Godot games1229Not measuredOnly Grok 4.5 measured
Design Arena: agents building games1235Not measuredOnly Grok 4.5 measured
Design Arena: HTML slides1204Not measuredOnly Grok 4.5 measured
Design Arena: PowerPoint slides1232Not measuredOnly Grok 4.5 measured
LiveBench research85.9%Not measuredOnly Grok 4.5 measured
LiveBench browser44.0%Not measuredOnly Grok 4.5 measured
LiveBench long agentic27.9%Not measuredOnly Grok 4.5 measured
SREBench (Vals AI)0.8%Not measuredOnly Grok 4.5 measured
Finance Agent (Vals AI)48.4%36.7%Grok 4.5
Tax Agent Bench (Vals AI)24.9%7.8%Grok 4.5
Legal Research Bench (Vals AI)38.0%9.1%Grok 4.5
Excel Modeling Benchmark (Vals AI)52.9%55.1%MiMo-V2.5
Time Horizon Index (Vals AI)6.3%Not measuredOnly Grok 4.5 measured
τ²-Bench TelecomNot measured90.6%Only MiMo-V2.5 measured
APEX-Agents56.2%Not measuredOnly Grok 4.5 measured
PostTrainBench23.4%Not measuredOnly Grok 4.5 measured
Vending-Bench 2$3,887Not measuredOnly Grok 4.5 measured
LMArena Search1216Not measuredOnly Grok 4.5 measured
Neither model has a result on the other 16 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index81
Reasoning benchmarks for Grok 4.5 and MiMo-V2.5
BenchmarkGrok 4.5MiMo-V2.5Leader
GPQA Diamond92.9%81.6%Grok 4.5
LiveBench reasoning87.2%Not measuredOnly Grok 4.5 measured
MMMU Pro (Vals AI)61.8%80.0%MiMo-V2.5
GPQA Diamond (Artificial Analysis run)93.1%84.9%Grok 4.5
Humanity's Last Exam (Artificial Analysis run)42.7%27.2%Grok 4.5
GPQA Diamond (Epoch AI run)93.4%Not measuredOnly Grok 4.5 measured
ARC-AGI87.2%Not measuredOnly Grok 4.5 measured
ARC-AGI-252.6%Not measuredOnly Grok 4.5 measured
SimpleBench70.0%Not measuredOnly Grok 4.5 measured
Chess puzzles35.0%Not measuredOnly Grok 4.5 measured
LMArena Text: hard prompts14621450Grok 4.5
LMArena Text: expert14651459Grok 4.5
LMArena Text: hard prompts in English14641461Grok 4.5
LMArena Vision: diagrams13041269Grok 4.5
LMArena Vision: homework12951262Grok 4.5
Neither model has a result on the other 6 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index30
Math benchmarks for Grok 4.5 and MiMo-V2.5
BenchmarkGrok 4.5MiMo-V2.5Leader
OTIS Mock AIME97.8%Not measuredOnly Grok 4.5 measured
FrontierMath, tiers 1 to 357.2%Not measuredOnly Grok 4.5 measured
FrontierMath, tier 424.4%Not measuredOnly Grok 4.5 measured
ProofBench31.0%16.0%Grok 4.5
LMArena Text: math14601436Grok 4.5
LMArena Text: mathematical work14821451Grok 4.5
Neither model has a result on the other 7 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index91
Knowledge benchmarks for Grok 4.5 and MiMo-V2.5
BenchmarkGrok 4.5MiMo-V2.5Leader
MMLU Pro89.2%82.9%Grok 4.5
LegalBench (Vals AI)86.0%78.9%Grok 4.5
CorpFin (Vals AI)67.4%59.9%Grok 4.5
TaxEval (Vals AI)71.7%71.8%MiMo-V2.5
MortgageTax (Vals AI)61.8%59.3%Grok 4.5
MedCode (Vals AI)43.3%31.9%Grok 4.5
SimpleQA Verified48.3%Not measuredOnly Grok 4.5 measured
LMArena Text: science14641440Grok 4.5
LMArena Text: medicine and healthcare14401432Grok 4.5
LMArena Text: legal and government14631425Grok 4.5
LMArena Text: business and finance14411432Grok 4.5
Neither model has a result on the other 6 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index80
Web design benchmarks for Grok 4.5 and MiMo-V2.5
BenchmarkGrok 4.5MiMo-V2.5Leader
Design Arena: websites12811266Grok 4.5
Design Arena: UI components12791257Grok 4.5
Design Arena: game dev12631249Grok 4.5
Design Arena: data viz12671249Grok 4.5
Design Arena: 3D12571216Grok 4.5
Design Arena: all code categories12761260Grok 4.5
Design Arena: SVG12151178Grok 4.5
Design Arena: ASCII art12571141Grok 4.5
LiveBench design78.1%Not measuredOnly Grok 4.5 measured
GeneralLike-for-likePublic boards, shown as published, kept out of the index261
General benchmarks for Grok 4.5 and MiMo-V2.5
BenchmarkGrok 4.5MiMo-V2.5Leader
LiveBench data analysis73.0%Not measuredOnly Grok 4.5 measured
SAGE (Vals AI)35.0%43.3%MiMo-V2.5
MedScribe (Vals AI)86.9%72.2%Grok 4.5
Vals Multimodal Index63.4%52.8%Grok 4.5
IFBenchNot measured67.1%Only MiMo-V2.5 measured
Long context reasoning (AA-LCR)79.3%73.0%Grok 4.5
DTBench96.5%Not measuredOnly Grok 4.5 measured
LMCA45.2%Not measuredOnly Grok 4.5 measured
Surface Evolver Bench74.4%Not measuredOnly Grok 4.5 measured
Furniture assembly22.5%Not measuredOnly Grok 4.5 measured
LMArena Text14481428Grok 4.5
LMArena Text: instruction following14451435Grok 4.5
LMArena Text: creative writing14421393Grok 4.5
LMArena Vision12871247Grok 4.5
LMArena Text: English14531447Grok 4.5
LMArena Text: non-English14401405Grok 4.5
LMArena Text: Chinese14951469Grok 4.5
LMArena Text: German14481421Grok 4.5
LMArena Text: Japanese14301306Grok 4.5
LMArena Text: Korean14061361Grok 4.5
LMArena Text: French14571448Grok 4.5
LMArena Text: Spanish14511414Grok 4.5
LMArena Text: Russian14491396Grok 4.5
LMArena Text: Polish14321413Grok 4.5
LMArena Text: multi-turn14561445Grok 4.5
LMArena Text: longer queries14621445Grok 4.5
LMArena Text: writing and language14451409Grok 4.5
LMArena Text: entertainment and media14341399Grok 4.5
LMArena Vision: English12851259Grok 4.5
LMArena Vision: Chinese13451326Grok 4.5
LMArena Vision: reading text12931256Grok 4.5
LMArena Vision: humor12761226Grok 4.5
LMArena Vision: recognizing thingsNot measured1189Only MiMo-V2.5 measured
LMArena Vision: creative writing12931230Grok 4.5
LMArena Document1452Not measuredOnly Grok 4.5 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 Grok 4.5 and MiMo-V2.5
BenchmarkGrok 4.5MiMo-V2.5Leader
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

Grok 4.5$0.0050Fits in one request
MiMo-V2.5$0.0002Fits in one request

MiMo-V2.5 costs less

Repository review

50K input + 3K output tokens

Grok 4.5$0.118Fits in one request
MiMo-V2.5$0.0067Fits in one request

MiMo-V2.5 costs less

Agent loop

200K input + 10K output tokens

Grok 4.5$0.460Fits in one request
MiMo-V2.5$0.026Fits in one request

MiMo-V2.5 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 Grok 4.5 and MiMo-V2.5
SpecificationSpaceXAI Grok 4.5Xiaomi MiMo-V2.5
MakerxAIXiaomi
WeightsNo public weights listedOpen weights
ReleasedJuly 8, 2026April 22, 2026
Context window500K1.05M
Input $/Mtok$2.00$0.119
Output $/Mtok$6.00$0.238
Coding index75.534.0
Independent measurements behind it21
Cost per solved task$0.62$0.01
SourceOpenRouter listingOpenRouter listing

Questions

Is Grok 4.5 better than MiMo-V2.5 for coding?

Pick Grok 4.5 for coding: it scores 75.5 on the Vibe Coding Index against 34.0 for MiMo-V2.5. Pick MiMo-V2.5 when output tokens cost 25.2x less, or when a solved task costs $0.01 against $0.62.

Which is cheaper, Grok 4.5 or MiMo-V2.5?

Grok 4.5 costs $2.00 per million input tokens and $6.00 per million output tokens; MiMo-V2.5 costs $0.119 and $0.238. Per solved SWE-bench Verified task, Grok 4.5 costs $0.62 and MiMo-V2.5 $0.01.

Which has the bigger context window, Grok 4.5 or MiMo-V2.5?

MiMo-V2.5, with 1.05M tokens against 500K.

How many benchmarks is the Grok 4.5 vs MiMo-V2.5 comparison based on?

76 benchmarks have a result for both models as of October 21, 2026. Grok 4.5's coding index rests on 2 independent benchmarks and MiMo-V2.5's on 1. Where only one model has a result, the page says so instead of naming a winner.

Compare next