LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

Kimi K2.5 vs Mistral Medium 3.5

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 Kimi K2.5 Replace Mistral Medium 3.5
Kimi K2.5Moonshot AI
Coding
58.5
Overall
23.5
Output $/M
$2.50
70
Mistral Medium 3.5Mistral AI
Coding
9.2
Overall
14.2
Output $/M
$7.50
2
vs boards won of 72 shared

Decision reading

Pick Kimi K2.5 for coding: it scores 58.5 on the Vibe Coding Index against 9.2 for Mistral Medium 3.5.

Pick Mistral Medium 3.5 when what matters is AA Coding Index or LMArena Text: German, where it leads.

Model AKimi K2.5Moonshot AI58.5/1002 independent benchmarks$0.24 per solved taskModel BMistral Medium 3.5Mistral AI9.2/1001 independent benchmark$6.02 per solved task
Shared results
72
Kimi K2.5 only
26
Mistral Medium 3.5 only
8
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
Kimi K2.5 58.5 from 2 benchmarks, Mistral Medium 3.5 9.2 from 1 benchmark.
Widest category
Web design, 16.4 points apart on average over 8 shared benchmarks.
Widest single gap
GPQA Diamond: Kimi K2.5 84.1% against Mistral Medium 3.5 34.9%, 49.2 points apart on a 0 to 100 scale.
Output price
Kimi K2.5 $2.50 against Mistral Medium 3.5 $7.50 per million output tokens, 3.0x apart.
Context window
Both take 262K tokens.

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
CategoryKimi K2.5BarsMistral Medium 3.5Ties
Coding111
Intelligence20
Agentic work40
Reasoning80
Math20
Knowledge100
Web design80
General251
All shared boards702
CodingLike-for-likeFour of these boards feed the Vibe Coding Index111
Coding benchmarks for Kimi K2.5 and Mistral Medium 3.5
BenchmarkKimi K2.5Mistral Medium 3.5Leader
SWE-bench Verified70.8%Not measuredOnly Kimi K2.5 measured
SWE-bench Multilingual67.3%Not measuredOnly Kimi K2.5 measured
LMArena WebDev14371264Kimi K2.5
AA Coding Index46.846.9Mistral Medium 3.5
LiveCodeBench (Vals AI)83.9%Not measuredOnly Kimi K2.5 measured
SWE-bench Verified (Vals AI)70.0%66.4%Kimi K2.5
Vibe Code Bench (Vals AI)17.5%2.9%Kimi K2.5
Code migration (Vals AI)7.0%5.1%Kimi K2.5
Terminal-Bench 2.1 (Vals AI)42.0%39.0%Kimi K2.5
SciCodeNot measured40.2%Only Mistral Medium 3.5 measured
Terminal-Bench Hard (Artificial Analysis run)34.8%33.3%Kimi K2.5
Terminal-Bench43.2%Not measuredOnly Kimi K2.5 measured
WeirdML45.6%Not measuredOnly Kimi K2.5 measured
ALE-Bench821.6764.0Kimi K2.5
LMArena Text: coding14831460Kimi K2.5
LMArena Text: software and IT14711458Kimi K2.5
LMArena WebDev: React14291254Kimi K2.5
LMArena WebDev: HTML14261308Kimi K2.5
LMArena WebDev: from an image1439Not measuredOnly Kimi K2.5 measured
Neither model has a result on the other 19 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
Intelligence benchmarks for Kimi K2.5 and Mistral Medium 3.5
BenchmarkKimi K2.5Mistral Medium 3.5Leader
AA Intelligence Index23.514.2Kimi K2.5
Epoch Capabilities Index148.1141.4Kimi K2.5
Agentic workLike-for-likePublic boards, shown as published, kept out of the index40
Agentic work benchmarks for Kimi K2.5 and Mistral Medium 3.5
BenchmarkKimi K2.5Mistral Medium 3.5Leader
AA Agentic IndexNot measured7.0Only Mistral Medium 3.5 measured
Design Arena: agents building web apps1136Not measuredOnly Kimi K2.5 measured
Design Arena: agents building full-stack apps1114Not measuredOnly Kimi K2.5 measured
Design Arena: agents building mobile apps1147Not measuredOnly Kimi K2.5 measured
Design Arena: agents building Android apps1124Not measuredOnly Kimi K2.5 measured
Design Arena: agents building Godot games1219Not measuredOnly Kimi K2.5 measured
Finance Agent (Vals AI)35.8%32.1%Kimi K2.5
Tax Agent Bench (Vals AI)Not measured4.5%Only Mistral Medium 3.5 measured
Legal Research Bench (Vals AI)15.9%9.1%Kimi K2.5
Excel Modeling Benchmark (Vals AI)28.5%11.2%Kimi K2.5
τ²-Bench Telecom95.9%94.2%Kimi K2.5
OSWorld63.3%Not measuredOnly Kimi K2.5 measured
Vending-Bench 2$1,198Not measuredOnly Kimi K2.5 measured
Neither model has a result on the other 26 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index80
Reasoning benchmarks for Kimi K2.5 and Mistral Medium 3.5
BenchmarkKimi K2.5Mistral Medium 3.5Leader
GPQA Diamond84.1%34.9%Kimi K2.5
MMMU Pro (Vals AI)84.3%Not measuredOnly Kimi K2.5 measured
GPQA Diamond (Artificial Analysis run)87.9%74.8%Kimi K2.5
Humanity's Last Exam (Artificial Analysis run)30.7%13.8%Kimi K2.5
ARC-AGI65.3%Not measuredOnly Kimi K2.5 measured
ARC-AGI-211.8%Not measuredOnly Kimi K2.5 measured
SimpleBench46.8%Not measuredOnly Kimi K2.5 measured
LMArena Text: hard prompts14531436Kimi K2.5
LMArena Text: expert14661432Kimi K2.5
LMArena Text: hard prompts in English14601445Kimi K2.5
LMArena Vision: diagrams12821254Kimi K2.5
LMArena Vision: homework12971230Kimi K2.5
Neither model has a result on the other 9 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index20
Math benchmarks for Kimi K2.5 and Mistral Medium 3.5
BenchmarkKimi K2.5Mistral Medium 3.5Leader
AIME95.6%Not measuredOnly Kimi K2.5 measured
ProofBenchNot measured9.0%Only Mistral Medium 3.5 measured
LMArena Text: math14701431Kimi K2.5
LMArena Text: mathematical work14731440Kimi K2.5
Neither model has a result on the other 9 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index100
Knowledge benchmarks for Kimi K2.5 and Mistral Medium 3.5
BenchmarkKimi K2.5Mistral Medium 3.5Leader
MMLU Pro85.9%75.3%Kimi K2.5
CaseLaw (Vals AI)58.7%44.2%Kimi K2.5
CorpFin (Vals AI)68.3%58.8%Kimi K2.5
TaxEval (Vals AI)74.2%68.0%Kimi K2.5
MortgageTax (Vals AI)66.5%28.9%Kimi K2.5
MedQA (Vals AI)94.4%Not measuredOnly Kimi K2.5 measured
MedCode (Vals AI)39.3%33.8%Kimi K2.5
SimpleQA Verified34.3%Not measuredOnly Kimi K2.5 measured
LMArena Text: science14591428Kimi K2.5
LMArena Text: medicine and healthcare14481409Kimi K2.5
LMArena Text: legal and government14411437Kimi K2.5
LMArena Text: business and finance14271426Kimi K2.5
Neither model has a result on the other 5 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index80
Web design benchmarks for Kimi K2.5 and Mistral Medium 3.5
BenchmarkKimi K2.5Mistral Medium 3.5Leader
Design Arena: websites12491142Kimi K2.5
Design Arena: UI components12361079Kimi K2.5
Design Arena: game dev12071079Kimi K2.5
Design Arena: data viz12191120Kimi K2.5
Design Arena: 3D12051061Kimi K2.5
Design Arena: all code categories12361122Kimi K2.5
Design Arena: SVG1149949Kimi K2.5
Design Arena: ASCII art11651062Kimi K2.5
Neither model has a result on the other 1 board here
GeneralLike-for-likePublic boards, shown as published, kept out of the index251
General benchmarks for Kimi K2.5 and Mistral Medium 3.5
BenchmarkKimi K2.5Mistral Medium 3.5Leader
SAGE (Vals AI)49.9%37.6%Kimi K2.5
MedScribe (Vals AI)76.4%67.7%Kimi K2.5
Vals Multimodal IndexNot measured34.8%Only Mistral Medium 3.5 measured
IFBench70.2%68.8%Kimi K2.5
Long context reasoning (AA-LCR)78.0%69.3%Kimi K2.5
DTBenchNot measured75.5%Only Mistral Medium 3.5 measured
LMCANot measured26.1%Only Mistral Medium 3.5 measured
Fiction.LiveBench86.1%Not measuredOnly Kimi K2.5 measured
CL-bench19.3%Not measuredOnly Kimi K2.5 measured
Surface Evolver BenchNot measured26.9%Only Mistral Medium 3.5 measured
CL-bench Life13.2%Not measuredOnly Kimi K2.5 measured
LMArena Text14451421Kimi K2.5
LMArena Text: instruction following14311415Kimi K2.5
LMArena Text: creative writing14221374Kimi K2.5
LMArena Vision12691222Kimi K2.5
LMArena Text: English14531433Kimi K2.5
LMArena Text: non-English14331404Kimi K2.5
LMArena Text: Chinese14941444Kimi K2.5
LMArena Text: German14411453Mistral Medium 3.5
LMArena Text: Japanese1423Not measuredOnly Kimi K2.5 measured
LMArena Text: Korean14121384Kimi K2.5
LMArena Text: French14541448Kimi K2.5
LMArena Text: Spanish14501411Kimi K2.5
LMArena Text: Russian14351395Kimi K2.5
LMArena Text: Polish14451430Kimi K2.5
LMArena Text: multi-turn14441423Kimi K2.5
LMArena Text: longer queries14451415Kimi K2.5
LMArena Text: writing and language14251384Kimi K2.5
LMArena Text: entertainment and media14171382Kimi K2.5
LMArena Vision: English12741240Kimi K2.5
LMArena Vision: Chinese13281244Kimi K2.5
LMArena Vision: reading text12761233Kimi K2.5
LMArena Vision: humor12841216Kimi K2.5
LMArena Vision: captioning1257Not measuredOnly Kimi K2.5 measured
LMArena Vision: recognizing things1285Not measuredOnly Kimi K2.5 measured
LMArena Vision: creative writing12691198Kimi K2.5
Neither model has a result on the other 10 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Kimi K2.5 and Mistral Medium 3.5
BenchmarkKimi K2.5Mistral Medium 3.5Leader
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

Kimi K2.5$0.0018Fits in one request
Mistral Medium 3.5$0.0053Fits in one request

Kimi K2.5 costs less

Repository review

50K input + 3K output tokens

Kimi K2.5$0.033Fits in one request
Mistral Medium 3.5$0.098Fits in one request

Kimi K2.5 costs less

Agent loop

200K input + 10K output tokens

Kimi K2.5$0.125Fits in one request
Mistral Medium 3.5$0.375Fits in one request

Kimi K2.5 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 Kimi K2.5 and Mistral Medium 3.5
SpecificationMoonshotAI Kimi K2.5Mistral Medium 3.5
MakerMoonshot AIMistral AI
WeightsOpen weightsNo public weights listed
ReleasedJanuary 27, 2026April 30, 2026
Context window262K262K
Input $/Mtok$0.50$1.50
Output $/Mtok$2.50$7.50
Coding index58.59.2
Independent measurements behind it21
Cost per solved task$0.24$6.02
SourceOpenRouter listingOpenRouter listing

Questions

Is Kimi K2.5 better than Mistral Medium 3.5 for coding?

Pick Kimi K2.5 for coding: it scores 58.5 on the Vibe Coding Index against 9.2 for Mistral Medium 3.5. Pick Mistral Medium 3.5 when what matters is AA Coding Index or LMArena Text: German, where it leads.

Which is cheaper, Kimi K2.5 or Mistral Medium 3.5?

Kimi K2.5 costs $0.50 per million input tokens and $2.50 per million output tokens; Mistral Medium 3.5 costs $1.50 and $7.50. Per solved SWE-bench Verified task, Kimi K2.5 costs $0.24 and Mistral Medium 3.5 $6.02.

Which has the bigger context window, Kimi K2.5 or Mistral Medium 3.5?

Neither: both take 262K tokens.

How many benchmarks is the Kimi K2.5 vs Mistral Medium 3.5 comparison based on?

72 benchmarks have a result for both models as of October 21, 2026. Kimi K2.5's coding index rests on 2 independent benchmarks and Mistral Medium 3.5's on 1. Where only one model has a result, the page says so instead of naming a winner.

Compare next