LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

MiniMax M2.7 vs Qwen3.5-27B

Updated October 21, 2026. 40 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.

Replace MiniMax M2.7 Replace Qwen3.5-27B
MiniMax M2.7MiniMax
Coding
28.4
Overall
22.8
Output $/M
$1.20
23
Qwen3.5-27BQwen
Coding
22.6
Overall
22.9
Output $/M
$2.16
17
vs boards won of 40 shared

Decision reading

Pick MiniMax M2.7 for coding: it scores 28.4 on the Vibe Coding Index against 22.6 for Qwen3.5-27B.

Pick Qwen3.5-27B when the context window is larger, 262K against 205K tokens, or when what matters is WeirdML or AA Intelligence Index, where it leads.

Model AMiniMax M2.7MiniMax28.4/1001 independent benchmark$0.64 per solved taskModel BQwen3.5-27BQwen22.6/1001 independent benchmarkNo cost per solved task
Shared results
40
MiniMax M2.7 only
34
Qwen3.5-27B only
13
Like-for-like categories
7 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
MiniMax M2.7 28.4 from 1 benchmark, Qwen3.5-27B 22.6 from 1 benchmark.
Widest category
Agentic work, 9.1 points apart on average over 1 shared benchmark.
Widest single gap
LMArena Text: Japanese: MiniMax M2.7 1260 against Qwen3.5-27B 1346, 14.3 points apart on a 0 to 100 scale.
Output price
MiniMax M2.7 $1.20 against Qwen3.5-27B $2.16 per million output tokens, 1.8x apart.
Context window
MiniMax M2.7 takes 205K tokens, Qwen3.5-27B 262K.

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
CategoryMiniMax M2.7BarsQwen3.5-27BTies
Coding71
Intelligence01
Agentic work01
Reasoning50
Math11
Knowledge13
General910
All shared boards2317
CodingLike-for-likeFour of these boards feed the Vibe Coding Index71
Coding benchmarks for MiniMax M2.7 and Qwen3.5-27B
BenchmarkMiniMax M2.7Qwen3.5-27BLeader
LMArena WebDev13991358MiniMax M2.7
AA Coding Index52.6Not measuredOnly MiniMax M2.7 measured
LiveCodeBench (Vals AI)79.9%Not measuredOnly MiniMax M2.7 measured
SWE-bench Verified (Vals AI)73.8%Not measuredOnly MiniMax M2.7 measured
Vibe Code Bench (Vals AI)11.9%Not measuredOnly MiniMax M2.7 measured
Code migration (Vals AI)8.7%Not measuredOnly MiniMax M2.7 measured
ProgramBench (Vals AI)0.0%Not measuredOnly MiniMax M2.7 measured
Terminal-Bench 2.1 (Vals AI)48.7%Not measuredOnly MiniMax M2.7 measured
SciCode50.1%Not measuredOnly MiniMax M2.7 measured
Terminal-Bench Hard (Artificial Analysis run)39.4%32.6%MiniMax M2.7
Terminal-Bench45.1%Not measuredOnly MiniMax M2.7 measured
WeirdML37.0%39.5%Qwen3.5-27B
ALE-Bench599.3349.4MiniMax M2.7
LMArena Text: coding14541427MiniMax M2.7
LMArena Text: software and IT14461431MiniMax M2.7
LMArena WebDev: React13881346MiniMax M2.7
LMArena WebDev: HTML14011386MiniMax M2.7
Neither model has a result on the other 21 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index01
Intelligence benchmarks for MiniMax M2.7 and Qwen3.5-27B
BenchmarkMiniMax M2.7Qwen3.5-27BLeader
AA Intelligence Index22.822.9Qwen3.5-27B
Epoch Capabilities Index145.8Not measuredOnly MiniMax M2.7 measured
Agentic workLike-for-likePublic boards, shown as published, kept out of the index01
Agentic work benchmarks for MiniMax M2.7 and Qwen3.5-27B
BenchmarkMiniMax M2.7Qwen3.5-27BLeader
AA Agentic Index15.3Not measuredOnly MiniMax M2.7 measured
Finance Agent (Vals AI)27.9%Not measuredOnly MiniMax M2.7 measured
Tax Agent Bench (Vals AI)6.9%Not measuredOnly MiniMax M2.7 measured
Legal Research Bench (Vals AI)10.6%Not measuredOnly MiniMax M2.7 measured
Excel Modeling Benchmark (Vals AI)28.4%Not measuredOnly MiniMax M2.7 measured
τ²-Bench Telecom84.8%93.9%Qwen3.5-27B
Vending-Bench 2Not measured$202Only Qwen3.5-27B measured
Neither model has a result on the other 32 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index50
Reasoning benchmarks for MiniMax M2.7 and Qwen3.5-27B
BenchmarkMiniMax M2.7Qwen3.5-27BLeader
GPQA Diamond86.6%Not measuredOnly MiniMax M2.7 measured
GPQA Diamond (Artificial Analysis run)87.4%85.8%MiniMax M2.7
Humanity's Last Exam (Artificial Analysis run)29.6%23.9%MiniMax M2.7
LMArena Text: hard prompts14221415MiniMax M2.7
LMArena Text: expert14431427MiniMax M2.7
LMArena Text: hard prompts in English14411429MiniMax M2.7
LMArena Vision: diagramsNot measured1249Only Qwen3.5-27B measured
LMArena Vision: homeworkNot measured1256Only Qwen3.5-27B measured
Neither model has a result on the other 13 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index11
Math benchmarks for MiniMax M2.7 and Qwen3.5-27B
BenchmarkMiniMax M2.7Qwen3.5-27BLeader
AIME91.0%Not measuredOnly MiniMax M2.7 measured
ProofBench3.0%Not measuredOnly MiniMax M2.7 measured
LMArena Text: math14211429Qwen3.5-27B
LMArena Text: mathematical work14371429MiniMax M2.7
Neither model has a result on the other 9 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index13
Knowledge benchmarks for MiniMax M2.7 and Qwen3.5-27B
BenchmarkMiniMax M2.7Qwen3.5-27BLeader
MMLU Pro80.4%Not measuredOnly MiniMax M2.7 measured
LegalBench (Vals AI)84.0%Not measuredOnly MiniMax M2.7 measured
CaseLaw (Vals AI)60.9%Not measuredOnly MiniMax M2.7 measured
CorpFin (Vals AI)61.2%Not measuredOnly MiniMax M2.7 measured
TaxEval (Vals AI)66.6%Not measuredOnly MiniMax M2.7 measured
MedCode (Vals AI)34.4%Not measuredOnly MiniMax M2.7 measured
LMArena Text: science14161437Qwen3.5-27B
LMArena Text: medicine and healthcare14091425Qwen3.5-27B
LMArena Text: legal and government14031405Qwen3.5-27B
LMArena Text: business and finance14071399MiniMax M2.7
Neither model has a result on the other 7 boards here
Web designDirectional onlyPublic boards, shown as published, kept out of the index8 boards, none shared
Web design benchmarks for MiniMax M2.7 and Qwen3.5-27B
BenchmarkMiniMax M2.7Qwen3.5-27BLeader
Design Arena: websites1243Not measuredOnly MiniMax M2.7 measured
Design Arena: UI components1210Not measuredOnly MiniMax M2.7 measured
Design Arena: game dev1210Not measuredOnly MiniMax M2.7 measured
Design Arena: data viz1233Not measuredOnly MiniMax M2.7 measured
Design Arena: 3D1195Not measuredOnly MiniMax M2.7 measured
Design Arena: all code categories1233Not measuredOnly MiniMax M2.7 measured
Design Arena: SVG1136Not measuredOnly MiniMax M2.7 measured
Design Arena: ASCII art1136Not measuredOnly MiniMax M2.7 measured
Neither model has a result on the other 1 board here
GeneralLike-for-likePublic boards, shown as published, kept out of the index910
General benchmarks for MiniMax M2.7 and Qwen3.5-27B
BenchmarkMiniMax M2.7Qwen3.5-27BLeader
MedScribe (Vals AI)79.9%Not measuredOnly MiniMax M2.7 measured
IFBench75.7%75.6%MiniMax M2.7
Long context reasoning (AA-LCR)78.3%77.7%MiniMax M2.7
DTBenchNot measured82.4%Only Qwen3.5-27B measured
LMCANot measured34.0%Only Qwen3.5-27B measured
ExploitBench13.3%Not measuredOnly MiniMax M2.7 measured
LMArena Text14051409Qwen3.5-27B
LMArena Text: instruction following14051393MiniMax M2.7
LMArena Text: creative writing13541362Qwen3.5-27B
LMArena VisionNot measured1241Only Qwen3.5-27B measured
LMArena Text: English14261425MiniMax M2.7
LMArena Text: non-English13821390Qwen3.5-27B
LMArena Text: Chinese14421478Qwen3.5-27B
LMArena Text: German13981392MiniMax M2.7
LMArena Text: Japanese12601346Qwen3.5-27B
LMArena Text: Korean13141359Qwen3.5-27B
LMArena Text: French14241410MiniMax M2.7
LMArena Text: Spanish14021407Qwen3.5-27B
LMArena Text: Russian13831390Qwen3.5-27B
LMArena Text: Polish13871375MiniMax M2.7
LMArena Text: multi-turn14121410MiniMax M2.7
LMArena Text: longer queries14181413MiniMax M2.7
LMArena Text: writing and language13761380Qwen3.5-27B
LMArena Text: entertainment and media13601363Qwen3.5-27B
LMArena Vision: EnglishNot measured1246Only Qwen3.5-27B measured
LMArena Vision: ChineseNot measured1301Only Qwen3.5-27B measured
LMArena Vision: reading textNot measured1246Only Qwen3.5-27B measured
LMArena Vision: humorNot measured1227Only Qwen3.5-27B measured
LMArena Vision: captioningNot measured1170Only Qwen3.5-27B measured
LMArena Vision: recognizing thingsNot measured1210Only Qwen3.5-27B measured
LMArena Vision: creative writingNot measured1224Only Qwen3.5-27B measured
Neither model has a result on the other 15 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for MiniMax M2.7 and Qwen3.5-27B
BenchmarkMiniMax M2.7Qwen3.5-27BLeader
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

MiniMax M2.7$0.0009Fits in one request
Qwen3.5-27B$0.0014Fits in one request

MiniMax M2.7 costs less

Repository review

50K input + 3K output tokens

MiniMax M2.7$0.019Fits in one request
Qwen3.5-27B$0.020Fits in one request

MiniMax M2.7 costs less

Agent loop

200K input + 10K output tokens

MiniMax M2.7$0.072Exceeds the context window
Qwen3.5-27B$0.076Fits in one request

MiniMax M2.7 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 MiniMax M2.7 and Qwen3.5-27B
SpecificationMiniMax M2.7Qwen3.5-27B
MakerMiniMaxQwen
WeightsOpen weightsOpen weights
ReleasedMarch 18, 2026February 25, 2026
Context window205K262K
Input $/Mtok$0.30$0.27
Output $/Mtok$1.20$2.16
Coding index28.422.6
Independent measurements behind it11
Cost per solved task$0.64Not computable
SourceOpenRouter listingOpenRouter listing

Questions

Is MiniMax M2.7 better than Qwen3.5-27B for coding?

Pick MiniMax M2.7 for coding: it scores 28.4 on the Vibe Coding Index against 22.6 for Qwen3.5-27B. Pick Qwen3.5-27B when the context window is larger, 262K against 205K tokens, or when what matters is WeirdML or AA Intelligence Index, where it leads.

Which is cheaper, MiniMax M2.7 or Qwen3.5-27B?

MiniMax M2.7 costs $0.30 per million input tokens and $1.20 per million output tokens; Qwen3.5-27B costs $0.27 and $2.16. Cost per solved task is known for MiniMax M2.7 only.

Which has the bigger context window, MiniMax M2.7 or Qwen3.5-27B?

Qwen3.5-27B, with 262K tokens against 205K.

How many benchmarks is the MiniMax M2.7 vs Qwen3.5-27B comparison based on?

40 benchmarks have a result for both models as of October 21, 2026. MiniMax M2.7's coding index rests on 1 independent benchmark and Qwen3.5-27B's on 1. Where only one model has a result, the page says so instead of naming a winner.

Compare next