Muse Ember vs MiniMax M2.7
Muse Ember leads MiniMax M2.7 on coding, 64.9 to 36.2 on the Vibe Coding Index, and Muse Ember costs 2.6x less per output token ($0.50 vs $1.32 per million). Updated November 14, 2026; 7 benchmarks have a result for both.
Muse Ember or MiniMax M2.7: which to pick
Pick Muse Ember for coding: it scores 64.9 on the Vibe Coding Index against 36.2 for MiniMax M2.7.
MiniMax M2.7 is neither cheaper nor larger in context, and leads no benchmark both models were measured on.
- Shared results
- 7
- Muse Ember only
- 8
- MiniMax M2.7 only
- 69
- Like-for-like categories
- 2 of 9
Only benchmarks that tested both models count toward a pick; a result only one model has is shown but not compared.
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.
Coding
Vibe Coding Index, six public coding benchmarks
Muse Ember
Muse Ember scores 64.9 against 36.2.
3 benchmarks behind each score
Agentic work
AA Agentic Index
No call
Only MiniMax M2.7 has a AA Agentic Index result, 15.3.
Directional only
Reasoning
GPQA Diamond
No call
Only MiniMax M2.7 has a GPQA Diamond result, 86.6%.
Directional only
Math
AIME
No call
Only MiniMax M2.7 has a AIME result, 91.0%.
Directional only
Knowledge
MMLU Pro
No call
Only MiniMax M2.7 has a MMLU Pro result, 80.4%.
Directional only
Web design
Design Arena: websites
No call
Only MiniMax M2.7 has a Design Arena: websites result, 1234.
Directional only
General
Vals Index
No call
Only Muse Ember has a Vals Index result, 60.5%.
Directional only
Long documents
Listed context window
Muse Ember
Muse Ember takes 1.05M tokens against 197K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Muse Ember
Muse Ember $0.0003, MiniMax M2.7 $0.0011 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
No call
Only MiniMax M2.7 has the number, $0.64.
Directional only
At a glance
- Coding index
- Muse Ember 64.9 from 3 benchmarks, MiniMax M2.7 36.2 from 3 benchmarks.
- Widest category
- Coding, 42.3 points apart on average over 5 shared benchmarks.
- Widest single gap
- Vibe Code Bench (Vals AI): Muse Ember 83.7% against MiniMax M2.7 11.9%, 71.8 points apart on a 0 to 100 scale.
- Output price
- Muse Ember $0.50 against MiniMax M2.7 $1.32 per million output tokens, 2.6x apart.
- Context window
- Muse Ember takes 1.05M tokens, MiniMax M2.7 197K.
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.
| Category | Muse Ember | Bars | MiniMax M2.7 | Ties |
|---|---|---|---|---|
| Coding | 5 | 0 | ||
| Agentic work | 2 | 0 | ||
| All shared boards | 7 | 0 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index50
| Benchmark | Muse Ember | MiniMax M2.7 | Leader |
|---|---|---|---|
| LMArena WebDev | 1609 | 1398 | Muse Ember |
| AA Coding Index | Not measured | 52.6 | Only MiniMax M2.7 measured |
| LiveCodeBench (Vals AI) | Not measured | 79.9% | Only MiniMax M2.7 measured |
| SWE-bench Verified (Vals AI) | Not measured | 73.8% | Only MiniMax M2.7 measured |
| Terminal-Bench 4 (Vals AI) | 39.4% | Not measured | Only Muse Ember measured |
| Vibe Code Bench (Vals AI) | 83.7% | 11.9% | Muse Ember |
| Code migration (Vals AI) | Not measured | 8.7% | Only MiniMax M2.7 measured |
| ProgramBench (Vals AI) | Not measured | 0.0% | Only MiniMax M2.7 measured |
| Terminal-Bench 2.1 (Vals AI) | Not measured | 48.7% | Only MiniMax M2.7 measured |
| SciCode | Not measured | 50.1% | Only MiniMax M2.7 measured |
| Terminal-Bench Hard (Artificial Analysis run) | Not measured | 39.4% | Only MiniMax M2.7 measured |
| Terminal-Bench 4.0 (Artificial Analysis run) | 40.9% | 0.0% | Muse Ember |
| FrontierCode 2.0 | 27.2% | Not measured | Only Muse Ember measured |
| Terminal-Bench | Not measured | 45.1% | Only MiniMax M2.7 measured |
| WeirdML | Not measured | 37.0% | Only MiniMax M2.7 measured |
| ALE-Bench | Not measured | 599.3 | Only MiniMax M2.7 measured |
| LMArena Text: coding | Not measured | 1453 | Only MiniMax M2.7 measured |
| LMArena Text: software and IT | Not measured | 1446 | Only MiniMax M2.7 measured |
| LMArena WebDev: React | 1616 | 1386 | Muse Ember |
| LMArena WebDev: HTML | 1621 | 1401 | Muse Ember |
| Terminal-Bench 2.0 | Not measured | 45.1% | Only MiniMax M2.7 measured |
| Neither model has a result on the other 31 boards here | |||
IntelligenceNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Muse Ember | MiniMax M2.7 | Leader |
|---|---|---|---|
| Neither model has a result on 2 boards here | |||
Agentic workLike-for-likePublic boards, shown as published, kept out of the index20
| Benchmark | Muse Ember | MiniMax M2.7 | Leader |
|---|---|---|---|
| AA Agentic Index | Not measured | 15.3 | Only MiniMax M2.7 measured |
| Finance Agent (Vals AI) | Not measured | 27.9% | Only MiniMax M2.7 measured |
| Tax Agent Bench (Vals AI) | 38.2% | 6.9% | Muse Ember |
| Legal Research Bench (Vals AI) | 49.0% | 10.6% | Muse Ember |
| Harvey's Legal Agent Benchmark (Vals AI) | 25.0% | Not measured | Only Muse Ember measured |
| Excel Modeling Benchmark (Vals AI) | Not measured | 28.4% | Only MiniMax M2.7 measured |
| τ²-Bench Telecom | Not measured | 84.8% | Only MiniMax M2.7 measured |
| APEX-Agents | 64.3% | Not measured | Only Muse Ember measured |
| Toolathlon-Verified | Not measured | 47.5% | Only MiniMax M2.7 measured |
| Neither model has a result on the other 35 boards here | |||
ReasoningDirectional onlyPublic boards, shown as published, kept out of the index8 boards, none shared
| Benchmark | Muse Ember | MiniMax M2.7 | Leader |
|---|---|---|---|
| GPQA Diamond | Not measured | 86.6% | Only MiniMax M2.7 measured |
| GPQA Diamond (Artificial Analysis run) | Not measured | 87.4% | Only MiniMax M2.7 measured |
| Humanity's Last Exam (Artificial Analysis run) | Not measured | 29.6% | Only MiniMax M2.7 measured |
| HLE-Diamond (Artificial Analysis run) | 25.1% | Not measured | Only Muse Ember measured |
| CritPt v2, hard subset (Artificial Analysis run) | 39.0% | Not measured | Only Muse Ember measured |
| LMArena Text: hard prompts | Not measured | 1422 | Only MiniMax M2.7 measured |
| LMArena Text: expert | Not measured | 1443 | Only MiniMax M2.7 measured |
| LMArena Text: hard prompts in English | Not measured | 1440 | Only MiniMax M2.7 measured |
| Neither model has a result on the other 18 boards here | |||
MathDirectional onlyPublic boards, shown as published, kept out of the index4 boards, none shared
| Benchmark | Muse Ember | MiniMax M2.7 | Leader |
|---|---|---|---|
| AIME | Not measured | 91.0% | Only MiniMax M2.7 measured |
| ProofBench | Not measured | 3.0% | Only MiniMax M2.7 measured |
| LMArena Text: math | Not measured | 1422 | Only MiniMax M2.7 measured |
| LMArena Text: mathematical work | Not measured | 1437 | Only MiniMax M2.7 measured |
| Neither model has a result on the other 10 boards here | |||
KnowledgeDirectional onlyPublic boards, shown as published, kept out of the index11 boards, none shared
| Benchmark | Muse Ember | MiniMax M2.7 | Leader |
|---|---|---|---|
| MMLU Pro | Not measured | 80.4% | Only MiniMax M2.7 measured |
| LegalBench (Vals AI) | Not measured | 84.0% | Only MiniMax M2.7 measured |
| CaseLaw (Vals AI) | Not measured | 60.9% | Only MiniMax M2.7 measured |
| CorpFin (Vals AI) | Not measured | 61.2% | Only MiniMax M2.7 measured |
| TaxEval (Vals AI) | Not measured | 66.6% | Only MiniMax M2.7 measured |
| MedCode (Vals AI) | Not measured | 34.4% | Only MiniMax M2.7 measured |
| AA-Omniscience | 7.7 | Not measured | Only Muse Ember measured |
| LMArena Text: science | Not measured | 1416 | Only MiniMax M2.7 measured |
| LMArena Text: medicine and healthcare | Not measured | 1410 | Only MiniMax M2.7 measured |
| LMArena Text: legal and government | Not measured | 1402 | Only MiniMax M2.7 measured |
| LMArena Text: business and finance | Not measured | 1407 | Only MiniMax M2.7 measured |
| Neither model has a result on the other 10 boards here | |||
Web designDirectional onlyPublic boards, shown as published, kept out of the index8 boards, none shared
| Benchmark | Muse Ember | MiniMax M2.7 | Leader |
|---|---|---|---|
| Design Arena: websites | Not measured | 1234 | Only MiniMax M2.7 measured |
| Design Arena: UI components | Not measured | 1202 | Only MiniMax M2.7 measured |
| Design Arena: game dev | Not measured | 1198 | Only MiniMax M2.7 measured |
| Design Arena: data viz | Not measured | 1229 | Only MiniMax M2.7 measured |
| Design Arena: 3D | Not measured | 1175 | Only MiniMax M2.7 measured |
| Design Arena: all code categories | Not measured | 1223 | Only MiniMax M2.7 measured |
| Design Arena: SVG | Not measured | 1123 | Only MiniMax M2.7 measured |
| Design Arena: ASCII art | Not measured | 1129 | Only MiniMax M2.7 measured |
| Neither model has a result on the other 1 board here | |||
GeneralDirectional onlyPublic boards, shown as published, kept out of the index23 boards, none shared
| Benchmark | Muse Ember | MiniMax M2.7 | Leader |
|---|---|---|---|
| Vals Index | 60.5% | Not measured | Only Muse Ember measured |
| MedScribe (Vals AI) | Not measured | 79.9% | Only MiniMax M2.7 measured |
| IFBench | Not measured | 75.7% | Only MiniMax M2.7 measured |
| Long context reasoning (AA-LCR) | Not measured | 78.3% | Only MiniMax M2.7 measured |
| GDPval-AA | Not measured | 1015 | Only MiniMax M2.7 measured |
| ExploitBench | Not measured | 13.3% | Only MiniMax M2.7 measured |
| LMArena Text | Not measured | 1405 | Only MiniMax M2.7 measured |
| LMArena Text: instruction following | Not measured | 1405 | Only MiniMax M2.7 measured |
| LMArena Text: creative writing | Not measured | 1354 | Only MiniMax M2.7 measured |
| LMArena Text: English | Not measured | 1425 | Only MiniMax M2.7 measured |
| LMArena Text: non-English | Not measured | 1383 | Only MiniMax M2.7 measured |
| LMArena Text: Chinese | Not measured | 1439 | Only MiniMax M2.7 measured |
| LMArena Text: German | Not measured | 1398 | Only MiniMax M2.7 measured |
| LMArena Text: Japanese | Not measured | 1252 | Only MiniMax M2.7 measured |
| LMArena Text: Korean | Not measured | 1314 | Only MiniMax M2.7 measured |
| LMArena Text: French | Not measured | 1425 | Only MiniMax M2.7 measured |
| LMArena Text: Spanish | Not measured | 1406 | Only MiniMax M2.7 measured |
| LMArena Text: Russian | Not measured | 1383 | Only MiniMax M2.7 measured |
| LMArena Text: Polish | Not measured | 1386 | Only MiniMax M2.7 measured |
| LMArena Text: multi-turn | Not measured | 1412 | Only MiniMax M2.7 measured |
| LMArena Text: longer queries | Not measured | 1418 | Only MiniMax M2.7 measured |
| LMArena Text: writing and language | Not measured | 1376 | Only MiniMax M2.7 measured |
| LMArena Text: entertainment and media | Not measured | 1360 | Only MiniMax M2.7 measured |
| Neither model has a result on the other 26 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Muse Ember | MiniMax M2.7 | Leader |
|---|---|---|---|
| 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 Ember costs less
Repository review
50K input + 3K output tokens
Muse Ember costs less
Agent loop
200K input + 10K output tokens
Muse Ember 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.
| Specification | Meta Muse Ember | MiniMax M2.7 |
|---|---|---|
| Maker | Meta | MiniMax |
| Weights | No public weights listed | Open weights |
| Released | November 9, 2026 | March 18, 2026 |
| Context window | 1.05M | 197K |
| Input $/Mtok | $0.10 | $0.405 |
| Output $/Mtok | $0.50 | $1.32 |
| Coding index | 64.9 | 36.2 |
| Independent measurements behind it | 3 | 3 |
| Cost per solved task | Not computable | $0.64 |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Muse Ember better than MiniMax M2.7 for coding?
Pick Muse Ember for coding: it scores 64.9 on the Vibe Coding Index against 36.2 for MiniMax M2.7. MiniMax M2.7 is neither cheaper nor larger in context, and leads no benchmark both models were measured on.
Which is cheaper, Muse Ember or MiniMax M2.7?
Muse Ember costs $0.10 per million input tokens and $0.50 per million output tokens; MiniMax M2.7 costs $0.405 and $1.32. Cost per solved task is known for MiniMax M2.7 only.
Which has the bigger context window, Muse Ember or MiniMax M2.7?
Muse Ember, with 1.05M tokens against 197K.
How many benchmarks is the Muse Ember vs MiniMax M2.7 comparison based on?
7 benchmarks have a result for both models as of November 14, 2026. Muse Ember's coding index rests on 3 independent benchmarks and MiniMax M2.7's on 3. Where only one model has a result, the page says so instead of naming a winner.