Agentic
Like-for-like- MiniMax M3
- 42.0
- Supported · #112/154
- Muse Spark 1.2
- 61.0
- Supported · #16/154
- Basis
- BenchAlign lane · 9 vs 2 public rows
- Reading
- Muse Spark 1.2 leads · intervals overlap
MiniMax M3 vs Muse Spark 1.2
Decision reading
Muse Spark 1.2 has the higher public score estimate, 70.28 versus 61.25, but the 90% score intervals overlap. Treat that as a lead, not a settled winner.
3 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.
Updated September 18, 2026. Rank says Muse Spark 1.2 is ahead. Price, access, and your workload can each overturn that. Public scores include evidence status and uncertainty.
Both of these models will change. Get the price, version and retirement notices for the pair, each with its source. Follow model changes
Share or export
Which one for your work
Recommendations appear only when a shared evidence basis or an explicit operating constraint supports the call. Secondary and unsupported use cases stay disclosed below the initial list.
Agentic work
Tool use, computer use, and multi-step task completion
Muse Spark 1.2
Muse Spark 1.2 leads on the public agentic lane, 61 to 42, with Supported evidence for both models, although the 90% intervals overlap.
Confidence: limited
Chat turn cost
1K fresh input + 500 output tokens
MiniMax M3
MiniMax M3 has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-rates
Cache-heavy agent loop cost
200K cached + 20K fresh input + 10K output tokens
MiniMax M3
MiniMax M3 has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-rates
Show secondary and unsupported calls
Repository review cost
50K fresh input + 3K output tokens
MiniMax M3
MiniMax M3 has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-rates
Coding work
Code generation, repair, and software-engineering tasks
Not enough matched evidence
MiniMax M3 is scored on Estimated evidence for coding, so the reading is directional rather than like-for-like.
Confidence: limited
Long documents
Prompts that approach the documented context limit
No clear pick
The documented context windows are equal.
Confidence: documented
Which one for a specific job
Choose a job from the LLM Selector's task catalog to see the category row it rests on, under the same basis rules as the table below. A directional row stays directional; choosing a job never creates a winner.
The same task catalog as the LLM Selector. Each job names the evidence surface it rests on; nothing here adds a new score.
48.6MiniMax M360.0Muse Spark 1.2
Directional only · BenchAlign
Muse Spark 1.2 scores higher, but at least one score rests on Estimated evidence or a different benchmark set. Directional only, no winner.
Coding scores combine specific tasks and setups. Match the editor, harness, and effort to your workflow.
What is actually comparable
Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.
24 MiniMax M3 only
3 Shared
5 Muse Spark 1.2 only
- Shared results
- 3
- MiniMax M3 only
- 24
- Muse Spark 1.2 only
- 5
- Like-for-like categories
- 1 / 8
2 categories rest on Estimated evidence or different evidence sets. Those rows remain visible for coverage context but do not name a winner.
Category results, on a stated basis
Each row shows the public-lane category score for both models: the BenchAlign lane for agentic, coding, and knowledge, and the weighted public rows elsewhere, with the same rank each model page reports. A row is like-for-like only when both scores rest on Supported evidence or the same weighted set. Directional and not-comparable rows remain visible, but they never receive a winner.
Coding
Directional only- MiniMax M3
- 48.6
- Estimated · #69/154
- Muse Spark 1.2
- 60.0
- Supported · #21/154
- Basis
- BenchAlign lane · 10 vs 5 public rows
- Reading
- Directional only
Knowledge
Directional only- MiniMax M3
- 53.0
- Supported · #65/184
- Muse Spark 1.2
- 70.7
- Estimated · #10/184
- Basis
- BenchAlign lane · 2 vs 1 public rows
- Reading
- Directional only
Reasoning
Not comparable- MiniMax M3
- 78.0
- #4/20
- Muse Spark 1.2
- 75.3
- Unranked · 2 rankable rows
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Math
Not comparable- MiniMax M3
- Not ranked
- Muse Spark 1.2
- Not ranked
- Basis
- Provisional lane · 1 vs 0 weighted rows
- Reading
- Not comparable
Multilingual
Not comparable- MiniMax M3
- Not ranked
- Muse Spark 1.2
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Multimodal
Not comparable- MiniMax M3
- 52.1
- #34/48
- Muse Spark 1.2
- Not ranked
- Basis
- Provisional lane · 2 vs 0 weighted rows
- Reading
- Not comparable
Instruction following
Not comparable- MiniMax M3
- 92.4
- #4/124
- Muse Spark 1.2
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
| Category | MiniMax M3 | Muse Spark 1.2 | Basis | Reading |
|---|---|---|---|---|
| Agentic | 42.0Supported · #112/154 | 61.0Supported · #16/154 | Like-for-likeBenchAlign lane · 9 vs 2 public rows | Muse Spark 1.2 leads · intervals overlap |
| Coding | 48.6Estimated · #69/154 | 60.0Supported · #21/154 | Directional onlyBenchAlign lane · 10 vs 5 public rows | Directional only |
| Knowledge | 53.0Supported · #65/184 | 70.7Estimated · #10/184 | Directional onlyBenchAlign lane · 2 vs 1 public rows | Directional only |
| Reasoning | 78.0#4/20 | 75.3Unranked · 2 rankable rows | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Math | Not ranked | Not ranked | Not comparableProvisional lane · 1 vs 0 weighted rows | Not comparable |
| Multilingual | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Multimodal | 52.1#34/48 | Not ranked | Not comparableProvisional lane · 2 vs 0 weighted rows | Not comparable |
| Instruction following | 92.4#4/124 | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
Ranks count the models scored in each category’s lane, so the agentic, coding, and knowledge denominators (BenchAlign) differ from the provisional-lane categories. Unranked scores sit on the lane’s scale but never name a winner. Like-for-like rows within 0.5 points read as a practical tie.
Shape of the matched evidence
Only shared public evidence is shown. Sparse evidence stays a ruled list rather than being closed into a radar shape.
Too few matched category axes support a radar. The ruled list below shows only shared benchmark results; positions use each benchmark’s normalized display scale when available.
MMLU-Pro (Vals)
Knowledge
What each workload costs
Three fixed token mixes turn per-token rates into comparable decisions. Each scenario states context fit and whether cached input had to fall back to the published list-input rate.
Chat turn
1K fresh input + 500 output tokens
- MiniMax M3
- $0.0009
- Fits in one request
- Muse Spark 1.2
- $0.00338
- Fits in one request
MiniMax M3 has the lower modeled cost
Costs use the listed standard API rates.
Repository review
50K fresh input + 3K output tokens
- MiniMax M3
- $0.0186
- Fits in one request
- Muse Spark 1.2
- $0.07525
- Fits in one request
MiniMax M3 has the lower modeled cost
Costs use the listed standard API rates.
Cache-heavy agent loop
200K cached + 20K fresh input + 10K output tokens
- MiniMax M3
- $0.03
- Fits in one request
- Muse Spark 1.2
- $0.0975
- Fits in one request
MiniMax M3 has the lower modeled cost
Costs use the listed standard API rates.
Specification differences
Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.
SpecificationMiniMax M3Muse Spark 1.2
Context window
Maximum documented context; output-token limits may be lower.
MiniMax M3
1M
Muse Spark 1.2
API model ID
MiniMax M3
Not sourced
Muse Spark 1.2
muse-spark-1.2
Meta: Muse Spark 1.2 model pageCached-input rate
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
MiniMax M3
$0.06 per 1M cached input tokens
Muse Spark 1.2
$0.15 per 1M cached input tokens
Meta: Muse Spark 1.2 model pageDocumented inputs
MiniMax M3
Not sourced
Muse Spark 1.2
Not sourced
Documented outputs
MiniMax M3
Not sourced
Muse Spark 1.2
Not sourced
Provider availability
MiniMax M3
Not sourced
Muse Spark 1.2
Not sourced
Reasoning profile
MiniMax M3
Non-Reasoning
Muse Spark 1.2
Reasoning
Weight access
MiniMax M3
Open Weight
Muse Spark 1.2
Proprietary
License
MiniMax M3
Open Weight
Muse Spark 1.2
Proprietary
Release date
MiniMax M3
2026-06-01
Muse Spark 1.2
2026-08-05
If you already use one of these models
- Deployment change
- The models list different providers, so authentication, endpoint behavior, limits, and feature support may change.
- Quality signal
- Muse Spark 1.2 has the higher public score estimate, 70.28 versus 61.25, but the 90% score intervals overlap.
- Workload cost
- Repository review: $0.0186 vs $0.07525. Cache-heavy agent loop: $0.03 vs $0.0975.
- Context tradeoff
- Both models list 1M.
Run the same representative tasks against both endpoints before changing production traffic.
Benchmark evidence
The full public result ledger is available for audit without forcing a wide desktop table onto a phone.
Browse raw public benchmark evidence32 rows
Agentic
Terminal-Bench 2.0
MiniMax M366%SourceMuse Spark 1.2—Not directly comparable
BrowseComp
MiniMax M383.5%SourceMuse Spark 1.2—Not directly comparable
OSWorld-Verified
MiniMax M370.1%SourceMuse Spark 1.2—Not directly comparable
MCP Atlas
MiniMax M374.2%SourceMuse Spark 1.2—Not directly comparable
Claw-Eval
MiniMax M374.5%SourceMuse Spark 1.2—Not directly comparable
BankerToolBench
MiniMax M376.1%SourceMuse Spark 1.2—Not directly comparable
ResearchClawBench
MiniMax M319.8%SourceMuse Spark 1.2—Not directly comparable
OSWorld 2.0
MiniMax M34.6%SourceMuse Spark 1.2—Not directly comparable
Terminal-Bench 2.1 (Vals)
MiniMax M353.6%SourceMuse Spark 1.269.7%SourceMuse Spark 1.2 leads this result
Terminal-Bench 2.1
MiniMax M3—Muse Spark 1.282.9%SourceNot directly comparable
Coding
SWE-bench Verified
MiniMax M380.5%SourceMuse Spark 1.2—Not directly comparable
SWE-bench Pro
MiniMax M359%SourceMuse Spark 1.2—Not directly comparable
Terminal-Bench 2.0
MiniMax M366.0%SourceMuse Spark 1.2—Not directly comparable
NL2Repo
MiniMax M342.1%SourceMuse Spark 1.2—Not directly comparable
VIBE V2
MiniMax M350.1%SourceMuse Spark 1.2—Not directly comparable
SVG-Bench
MiniMax M363.7%SourceMuse Spark 1.2—Not directly comparable
KernelBench Hard
MiniMax M328.8%SourceMuse Spark 1.2—Not directly comparable
OpenHarmony Bench
MiniMax M348.4%SourceMuse Spark 1.2—Not directly comparable
LiveCodeBench (Vals)
MiniMax M382.2%SourceMuse Spark 1.2—Not directly comparable
SWE-bench (Vals)
MiniMax M375.0%SourceMuse Spark 1.286.6%SourceMuse Spark 1.2 leads this result
Terminal-Bench 2.1
MiniMax M3—Muse Spark 1.282.9%SourceNot directly comparable
DeepSWE
MiniMax M3—Muse Spark 1.259.3%SourceNot directly comparable
VulcanBench v3
MiniMax M3—Muse Spark 1.287.0%SourceNot directly comparable
FrontierSWE v2
MiniMax M3—Muse Spark 1.212.0%SourceNot directly comparable
Knowledge
Math
USAMO 2026
MiniMax M385.7%SourceMuse Spark 1.2—Not directly comparable
Multimodal
OfficeQA Pro
MiniMax M345.1%SourceMuse Spark 1.2—Not directly comparable
OmniDocBench 1.5
MiniMax M391.6%SourceMuse Spark 1.2—Not directly comparable
MMMU-Pro
MiniMax M378.1%SourceMuse Spark 1.2—Not directly comparable
VideoMMMU
MiniMax M384.6%SourceMuse Spark 1.2—Not directly comparable
Video-MME (with subtitle)
MiniMax M385.4%SourceMuse Spark 1.2—Not directly comparable
Questions
Which is better, MiniMax M3 or Muse Spark 1.2?
Muse Spark 1.2 has the higher public score estimate, 70.28 versus 61.25, but the 90% score intervals overlap. The higher estimate is not a decisive winner because the uncertainty ranges overlap.
Which is better for coding, MiniMax M3 or Muse Spark 1.2?
Muse Spark 1.2 scores higher for coding on the public lane, 60 to 48.6. MiniMax M3 is scored on Estimated evidence for coding, so the reading is directional rather than like-for-like. BenchLM does not name a winner for coding from a directional reading; read the shared benchmark rows directly and test the models on the same task set.
Which is better for agentic tasks, MiniMax M3 or Muse Spark 1.2?
Muse Spark 1.2 leads the public agentic tasks lane, 61 to 42, with Supported evidence for both models, although the 90% intervals overlap.
Which costs less, MiniMax M3 or Muse Spark 1.2?
For the stated presets, chat costs $0.0009 on MiniMax M3 and $0.00337 on Muse Spark 1.2; repository review costs $0.0186 and $0.07525; the cache-heavy agent loop costs $0.03 and $0.0975. Costs use the listed standard API rates.
Which has the larger context window, MiniMax M3 or Muse Spark 1.2?
Both models list the same context window, 1M.
Related comparisons
Last updated September 18, 2026
Watch MiniMax M3 vs Muse Spark 1.2
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 2,000+ readers.