LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

Muse Spark 1.1 vs MiniMax M2.7

Updated October 21, 2026. 63 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.

Replace Muse Spark 1.1 Replace MiniMax M2.7
Muse Spark 1.1Meta
Coding
48.9
Overall
33.7
Output $/M
$4.25
61
MiniMax M2.7MiniMax
Coding
28.4
Overall
22.8
Output $/M
$1.20
1
vs boards won of 63 shared, 1 tied

Decision reading

Pick Muse Spark 1.1 for coding: it scores 48.9 on the Vibe Coding Index against 28.4 for MiniMax M2.7.

Pick MiniMax M2.7 when output tokens cost 3.5x less, or when the benchmark that matters is Long context reasoning (AA-LCR), where it leads.

Model AMuse Spark 1.1Meta48.9/1001 independent benchmark$0.43 per solved taskModel BMiniMax M2.7MiniMax28.4/1001 independent benchmark$0.64 per solved task
Shared results
63
Muse Spark 1.1 only
33
MiniMax M2.7 only
11
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
Muse Spark 1.1 48.9 from 1 benchmark, MiniMax M2.7 28.4 from 1 benchmark.
Widest category
Agentic work, 23.8 points apart on average over 4 shared benchmarks.
Widest single gap
Vibe Code Bench (Vals AI): Muse Spark 1.1 72.2% against MiniMax M2.7 11.9%, 60.2 points apart on a 0 to 100 scale.
Output price
Muse Spark 1.1 $4.25 against MiniMax M2.7 $1.20 per million output tokens, 3.5x apart.
Context window
Muse Spark 1.1 takes 1.05M tokens, MiniMax M2.7 205K.

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
CategoryMuse Spark 1.1BarsMiniMax M2.7Ties
Coding1201 tie
Intelligence20
Agentic work40
Reasoning60
Math30
Knowledge80
Web design80
General181
All shared boards6111 tie
CodingLike-for-likeFour of these boards feed the Vibe Coding Index120
Coding benchmarks for Muse Spark 1.1 and MiniMax M2.7
BenchmarkMuse Spark 1.1MiniMax M2.7Leader
LMArena WebDev15421399Muse Spark 1.1
AA Coding Index71.352.6Muse Spark 1.1
LiveCodeBench (Vals AI)85.9%79.9%Muse Spark 1.1
SWE-bench Verified (Vals AI)82.0%73.8%Muse Spark 1.1
Vibe Code Bench (Vals AI)72.2%11.9%Muse Spark 1.1
Code migration (Vals AI)31.1%8.7%Muse Spark 1.1
ProgramBench (Vals AI)0.0%0.0%Tie
Terminal-Bench 2.1 (Vals AI)69.3%48.7%Muse Spark 1.1
SkillsBench (Vals AI)59.2%Not measuredOnly Muse Spark 1.1 measured
SciCode58.8%50.1%Muse Spark 1.1
Terminal-Bench Hard (Artificial Analysis run)Not measured39.4%Only MiniMax M2.7 measured
Terminal-BenchNot measured45.1%Only MiniMax M2.7 measured
WeirdMLNot measured37.0%Only MiniMax M2.7 measured
ALE-BenchNot measured599.3Only MiniMax M2.7 measured
LMArena Text: coding14981454Muse Spark 1.1
LMArena Text: software and IT14981446Muse Spark 1.1
LMArena WebDev: React15321388Muse Spark 1.1
LMArena WebDev: HTML15401401Muse Spark 1.1
LMArena WebDev: from an image1540Not measuredOnly Muse Spark 1.1 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 Muse Spark 1.1 and MiniMax M2.7
BenchmarkMuse Spark 1.1MiniMax M2.7Leader
AA Intelligence Index33.722.8Muse Spark 1.1
Epoch Capabilities Index154.1145.8Muse Spark 1.1
Agentic workLike-for-likePublic boards, shown as published, kept out of the index40
Agentic work benchmarks for Muse Spark 1.1 and MiniMax M2.7
BenchmarkMuse Spark 1.1MiniMax M2.7Leader
AA Agentic Index25.815.3Muse Spark 1.1
Design Arena: agents building web apps1207Not measuredOnly Muse Spark 1.1 measured
Design Arena: agents building full-stack apps1199Not measuredOnly Muse Spark 1.1 measured
Design Arena: agents building mobile apps1177Not measuredOnly Muse Spark 1.1 measured
Design Arena: agents building Android apps1184Not measuredOnly Muse Spark 1.1 measured
Design Arena: agents building Godot games1103Not measuredOnly Muse Spark 1.1 measured
Design Arena: agents building games1183Not measuredOnly Muse Spark 1.1 measured
Design Arena: HTML slides1194Not measuredOnly Muse Spark 1.1 measured
Design Arena: PowerPoint slides1156Not measuredOnly Muse Spark 1.1 measured
Design Arena: spreadsheets974Not measuredOnly Muse Spark 1.1 measured
Finance Agent (Vals AI)57.2%27.9%Muse Spark 1.1
Tax Agent Bench (Vals AI)Not measured6.9%Only MiniMax M2.7 measured
Legal Research Bench (Vals AI)38.0%10.6%Muse Spark 1.1
Excel Modeling Benchmark (Vals AI)56.4%28.4%Muse Spark 1.1
Time Horizon Index (Vals AI)1.7%Not measuredOnly Muse Spark 1.1 measured
τ²-Bench TelecomNot measured84.8%Only MiniMax M2.7 measured
APEX-Agents31.8%Not measuredOnly Muse Spark 1.1 measured
Vending-Bench 2$6,520Not measuredOnly Muse Spark 1.1 measured
Neither model has a result on the other 21 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index60
Reasoning benchmarks for Muse Spark 1.1 and MiniMax M2.7
BenchmarkMuse Spark 1.1MiniMax M2.7Leader
GPQA Diamond91.2%86.6%Muse Spark 1.1
MMMU Pro (Vals AI)86.6%Not measuredOnly Muse Spark 1.1 measured
GPQA Diamond (Artificial Analysis run)89.8%87.4%Muse Spark 1.1
Humanity's Last Exam (Artificial Analysis run)46.2%29.6%Muse Spark 1.1
Chess puzzles34.0%Not measuredOnly Muse Spark 1.1 measured
LMArena Text: hard prompts14851422Muse Spark 1.1
LMArena Text: expert14781443Muse Spark 1.1
LMArena Text: hard prompts in English14801441Muse Spark 1.1
LMArena Vision: diagrams1305Not measuredOnly Muse Spark 1.1 measured
LMArena Vision: homework1287Not measuredOnly Muse Spark 1.1 measured
Neither model has a result on the other 11 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index30
Math benchmarks for Muse Spark 1.1 and MiniMax M2.7
BenchmarkMuse Spark 1.1MiniMax M2.7Leader
AIMENot measured91.0%Only MiniMax M2.7 measured
ProofBench39.0%3.0%Muse Spark 1.1
LMArena Text: math14821421Muse Spark 1.1
LMArena Text: mathematical work14871437Muse Spark 1.1
Neither model has a result on the other 9 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index80
Knowledge benchmarks for Muse Spark 1.1 and MiniMax M2.7
BenchmarkMuse Spark 1.1MiniMax M2.7Leader
MMLU Pro88.7%80.4%Muse Spark 1.1
LegalBench (Vals AI)85.0%84.0%Muse Spark 1.1
CaseLaw (Vals AI)Not measured60.9%Only MiniMax M2.7 measured
CorpFin (Vals AI)71.3%61.2%Muse Spark 1.1
TaxEval (Vals AI)79.7%66.6%Muse Spark 1.1
MortgageTax (Vals AI)66.1%Not measuredOnly Muse Spark 1.1 measured
MedCode (Vals AI)Not measured34.4%Only MiniMax M2.7 measured
SimpleQA Verified57.8%Not measuredOnly Muse Spark 1.1 measured
LMArena Text: science14841416Muse Spark 1.1
LMArena Text: medicine and healthcare14791409Muse Spark 1.1
LMArena Text: legal and government14811403Muse Spark 1.1
LMArena Text: business and finance14701407Muse Spark 1.1
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 Muse Spark 1.1 and MiniMax M2.7
BenchmarkMuse Spark 1.1MiniMax M2.7Leader
Design Arena: websites12701243Muse Spark 1.1
Design Arena: UI components12901210Muse Spark 1.1
Design Arena: game dev12691210Muse Spark 1.1
Design Arena: data viz12681233Muse Spark 1.1
Design Arena: 3D12601195Muse Spark 1.1
Design Arena: all code categories12761233Muse Spark 1.1
Design Arena: SVG12341136Muse Spark 1.1
Design Arena: ASCII art12701136Muse Spark 1.1
Neither model has a result on the other 1 board here
GeneralLike-for-likePublic boards, shown as published, kept out of the index181
General benchmarks for Muse Spark 1.1 and MiniMax M2.7
BenchmarkMuse Spark 1.1MiniMax M2.7Leader
Public Benefits Bench (Vals AI)61.3%Not measuredOnly Muse Spark 1.1 measured
SAGE (Vals AI)45.6%Not measuredOnly Muse Spark 1.1 measured
MedScribe (Vals AI)88.9%79.9%Muse Spark 1.1
Vals Multimodal Index66.7%Not measuredOnly Muse Spark 1.1 measured
IFBenchNot measured75.7%Only MiniMax M2.7 measured
Long context reasoning (AA-LCR)77.7%78.3%MiniMax M2.7
DTBench94.4%Not measuredOnly Muse Spark 1.1 measured
LMCA49.9%Not measuredOnly Muse Spark 1.1 measured
Surface Evolver Bench52.5%Not measuredOnly Muse Spark 1.1 measured
ExploitBenchNot measured13.3%Only MiniMax M2.7 measured
LMArena Text14781405Muse Spark 1.1
LMArena Text: instruction following14571405Muse Spark 1.1
LMArena Text: creative writing14351354Muse Spark 1.1
LMArena Vision1293Not measuredOnly Muse Spark 1.1 measured
LMArena Text: English14781426Muse Spark 1.1
LMArena Text: non-English14711382Muse Spark 1.1
LMArena Text: Chinese15151442Muse Spark 1.1
LMArena Text: German14691398Muse Spark 1.1
LMArena Text: Japanese14501260Muse Spark 1.1
LMArena Text: Korean14591314Muse Spark 1.1
LMArena Text: French14921424Muse Spark 1.1
LMArena Text: Spanish14641402Muse Spark 1.1
LMArena Text: Russian14821383Muse Spark 1.1
LMArena Text: Polish14991387Muse Spark 1.1
LMArena Text: multi-turn14841412Muse Spark 1.1
LMArena Text: longer queries14621418Muse Spark 1.1
LMArena Text: writing and language14481376Muse Spark 1.1
LMArena Text: entertainment and media14381360Muse Spark 1.1
LMArena Vision: English1292Not measuredOnly Muse Spark 1.1 measured
LMArena Vision: Chinese1348Not measuredOnly Muse Spark 1.1 measured
LMArena Vision: reading text1301Not measuredOnly Muse Spark 1.1 measured
LMArena Vision: humor1288Not measuredOnly Muse Spark 1.1 measured
LMArena Vision: creative writing1253Not measuredOnly Muse Spark 1.1 measured
LMArena Document1465Not measuredOnly Muse Spark 1.1 measured
Neither model has a result on the other 12 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Muse Spark 1.1 and MiniMax M2.7
BenchmarkMuse Spark 1.1MiniMax M2.7Leader
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

Muse Spark 1.1$0.0034Fits in one request
MiniMax M2.7$0.0009Fits in one request

MiniMax M2.7 costs less

Repository review

50K input + 3K output tokens

Muse Spark 1.1$0.075Fits in one request
MiniMax M2.7$0.019Fits in one request

MiniMax M2.7 costs less

Agent loop

200K input + 10K output tokens

Muse Spark 1.1$0.292Fits in one request
MiniMax M2.7$0.072Exceeds the context window

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 Muse Spark 1.1 and MiniMax M2.7
SpecificationMeta Muse Spark 1.1MiniMax M2.7
MakerMetaMiniMax
WeightsNo public weights listedOpen weights
ReleasedJuly 16, 2026March 18, 2026
Context window1.05M205K
Input $/Mtok$1.25$0.30
Output $/Mtok$4.25$1.20
Coding index48.928.4
Independent measurements behind it11
Cost per solved task$0.43$0.64
SourceOpenRouter listingOpenRouter listing

Questions

Is Muse Spark 1.1 better than MiniMax M2.7 for coding?

Pick Muse Spark 1.1 for coding: it scores 48.9 on the Vibe Coding Index against 28.4 for MiniMax M2.7. Pick MiniMax M2.7 when output tokens cost 3.5x less, or when the benchmark that matters is Long context reasoning (AA-LCR), where it leads.

Which is cheaper, Muse Spark 1.1 or MiniMax M2.7?

Muse Spark 1.1 costs $1.25 per million input tokens and $4.25 per million output tokens; MiniMax M2.7 costs $0.30 and $1.20. Per solved SWE-bench Verified task, Muse Spark 1.1 costs $0.43 and MiniMax M2.7 $0.64.

Which has the bigger context window, Muse Spark 1.1 or MiniMax M2.7?

Muse Spark 1.1, with 1.05M tokens against 205K.

How many benchmarks is the Muse Spark 1.1 vs MiniMax M2.7 comparison based on?

63 benchmarks have a result for both models as of October 21, 2026. Muse Spark 1.1's coding index rests on 1 independent benchmark and MiniMax M2.7's on 1. Where only one model has a result, the page says so instead of naming a winner.

Compare next