Agentic
Like-for-like- MiniMax M3
- 42.0
- Supported · #109/152
- Qwen3.7 Plus
- 37.3
- Supported · #127/152
- Basis
- BenchAlign lane · 9 vs 11 public rows
- Reading
- MiniMax M3 leads · intervals overlap
MiniMax M3 vs Qwen3.7 Plus
Updated September 10, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.
Decision reading
Qwen3.7 Plus has the higher public score estimate, 61.57 versus 61.51, but the 90% score intervals overlap. Treat that as a lead, not a settled winner.
14 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.
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
MiniMax M3
MiniMax M3 leads on the public agentic lane, 42 to 37.3, with Supported evidence for both models, although the 90% intervals overlap.
Confidence: limited
Show secondary and unsupported calls
Coding work
Code generation, repair, and software-engineering tasks
Not enough matched evidence
MiniMax M3 and Qwen3.7 Plus are 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
Chat turn cost
1K fresh input + 500 output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
Confidence: listed-rates
Cache-heavy agent loop cost
200K cached + 20K fresh input + 10K output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
Confidence: listed-rates
Repository review cost
50K fresh input + 3K output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
Confidence: listed-rates
What is actually comparable
Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.
13 MiniMax M3 only
14 Shared
38 Qwen3.7 Plus only
- Shared results
- 14
- MiniMax M3 only
- 13
- Qwen3.7 Plus only
- 38
- Like-for-like categories
- 1 / 8
4 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
- 48.9
- Estimated · #66/151
- Qwen3.7 Plus
- 52.6
- Estimated · #47/151
- Basis
- BenchAlign lane · 10 vs 7 public rows
- Reading
- Directional only
Knowledge
Directional only- MiniMax M3
- 53.4
- Supported · #64/182
- Qwen3.7 Plus
- 56.3
- Estimated · #50/182
- Basis
- BenchAlign lane · 2 vs 7 public rows
- Reading
- Directional only
Multimodal
Directional only- MiniMax M3
- 52.2
- #34/48
- Qwen3.7 Plus
- 72.5
- #18/48
- Basis
- Provisional lane · 2 vs 2 weighted rows
- Reading
- Directional only
Instruction following
Directional only- MiniMax M3
- 93.7
- #4/121
- Qwen3.7 Plus
- 91.1
- #17/121
- Basis
- Provisional lane · 0 vs 1 weighted rows
- Reading
- Directional only
Reasoning
Not comparable- MiniMax M3
- 78.0
- Unranked · 2 rankable rows
- Qwen3.7 Plus
- 74.0
- Unranked · 3 rankable rows
- Basis
- Provisional lane · 0 vs 1 weighted rows
- Reading
- Not comparable
Math
Not comparable- MiniMax M3
- Not ranked
- Qwen3.7 Plus
- 78.3
- Unranked · 3 rankable rows
- Basis
- Provisional lane · 1 vs 1 weighted rows
- Reading
- Not comparable
Multilingual
Not comparable- MiniMax M3
- Not ranked
- Qwen3.7 Plus
- 78.9
- #3/12
- Basis
- Provisional lane · 0 vs 1 weighted rows
- Reading
- Not comparable
| Category | MiniMax M3 | Qwen3.7 Plus | Basis | Reading |
|---|---|---|---|---|
| Agentic | 42.0Supported · #109/152 | 37.3Supported · #127/152 | Like-for-likeBenchAlign lane · 9 vs 11 public rows | MiniMax M3 leads · intervals overlap |
| Coding | 48.9Estimated · #66/151 | 52.6Estimated · #47/151 | Directional onlyBenchAlign lane · 10 vs 7 public rows | Directional only |
| Knowledge | 53.4Supported · #64/182 | 56.3Estimated · #50/182 | Directional onlyBenchAlign lane · 2 vs 7 public rows | Directional only |
| Multimodal | 52.2#34/48 | 72.5#18/48 | Directional onlyProvisional lane · 2 vs 2 weighted rows | Directional only |
| Instruction following | 93.7#4/121 | 91.1#17/121 | Directional onlyProvisional lane · 0 vs 1 weighted rows | Directional only |
| Reasoning | 78.0Unranked · 2 rankable rows | 74.0Unranked · 3 rankable rows | Not comparableProvisional lane · 0 vs 1 weighted rows | Not comparable |
| Math | Not ranked | 78.3Unranked · 3 rankable rows | Not comparableProvisional lane · 1 vs 1 weighted rows | Not comparable |
| Multilingual | Not ranked | 78.9#3/12 | Not comparableProvisional lane · 0 vs 1 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.
Terminal-Bench 2.0
Agentic
OSWorld-Verified
Agentic
SWE-bench Verified
Coding
OSWorld 2.0
Agentic
SWE-bench Pro
Coding
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
- Qwen3.7 Plus
- API rate not published
- Fits in one request
Qwen3.7 Plus has no comparable published API token rate.
Repository review
50K fresh input + 3K output tokens
- MiniMax M3
- $0.0186
- Fits in one request
- Qwen3.7 Plus
- API rate not published
- Fits in one request
Qwen3.7 Plus has no comparable published API token rate.
Cache-heavy agent loop
200K cached + 20K fresh input + 10K output tokens
- MiniMax M3
- $0.03
- Fits in one request
- Qwen3.7 Plus
- API rate not published
- Fits in one request
- Cached-input rate unavailable
Qwen3.7 Plus has no comparable published API token rate.
Specification differences
Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.
SpecificationMiniMax M3Qwen3.7 Plus
Context window
Maximum documented context; output-token limits may be lower.
MiniMax M3
1M
Qwen3.7 Plus
1M
API model ID
MiniMax M3
Not sourced
Qwen3.7 Plus
Not sourced
Cached-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
Qwen3.7 Plus
No comparable hosted API rate
Documented inputs
MiniMax M3
Not sourced
Qwen3.7 Plus
Not sourced
Documented outputs
MiniMax M3
Not sourced
Qwen3.7 Plus
Not sourced
Provider availability
MiniMax M3
Not sourced
Qwen3.7 Plus
Not sourced
Reasoning profile
MiniMax M3
Non-Reasoning
Qwen3.7 Plus
Reasoning
Weight access
MiniMax M3
Open Weight
Qwen3.7 Plus
Proprietary
License
MiniMax M3
Open Weight
Qwen3.7 Plus
Proprietary
Release date
MiniMax M3
2026-06-01
Qwen3.7 Plus
2026-06-03
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
- Qwen3.7 Plus has the higher public score estimate, 61.57 versus 61.51, but the 90% score intervals overlap.
- Workload cost
- A complete comparable API-rate estimate is not available for both models.
- 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 evidence65 rows
Agentic
Terminal-Bench 2.0
MiniMax M366%SourceQwen3.7 Plus70.3%SourceQwen3.7 Plus leads this result
BrowseComp
MiniMax M383.5%SourceQwen3.7 Plus—Not directly comparable
OSWorld-Verified
MiniMax M370.1%SourceQwen3.7 Plus73.3%SourceQwen3.7 Plus leads this result
MCP Atlas
MiniMax M374.2%SourceQwen3.7 Plus73.2%SourceMiniMax M3 leads this result
Claw-Eval
MiniMax M374.5%SourceQwen3.7 Plus62.7%SourceMiniMax M3 leads this result
BankerToolBench
MiniMax M376.1%SourceQwen3.7 Plus—Not directly comparable
ResearchClawBench
MiniMax M319.8%SourceQwen3.7 Plus—Not directly comparable
OSWorld 2.0
Shared sourceMiniMax M34.6%Qwen3.7 Plus2.8%MiniMax M3 leads this result
Terminal-Bench 2.1 (Vals)
MiniMax M353.6%SourceQwen3.7 Plus52.8%SourceMiniMax M3 leads this result
QwenClawBench
MiniMax M3—Qwen3.7 Plus61.8%SourceNot directly comparable
BFCL v4
MiniMax M3—Qwen3.7 Plus72.9%SourceNot directly comparable
VITA-Bench
MiniMax M3—Qwen3.7 Plus45.6%SourceNot directly comparable
DeepPlanning
MiniMax M3—Qwen3.7 Plus62.3%SourceNot directly comparable
AndroidWorld
MiniMax M3—Qwen3.7 Plus81.0%SourceNot directly comparable
Coding
SWE-bench Verified
MiniMax M380.5%SourceQwen3.7 Plus77.7%SourceMiniMax M3 leads this result
SWE-bench Pro
MiniMax M359%SourceQwen3.7 Plus57.6%SourceMiniMax M3 leads this result
Terminal-Bench 2.0
MiniMax M366.0%SourceQwen3.7 Plus70.3%SourceQwen3.7 Plus leads this result
NL2Repo
MiniMax M342.1%SourceQwen3.7 Plus41.1%SourceMiniMax M3 leads this result
VIBE V2
MiniMax M350.1%SourceQwen3.7 Plus—Not directly comparable
SVG-Bench
MiniMax M363.7%SourceQwen3.7 Plus—Not directly comparable
KernelBench Hard
MiniMax M328.8%SourceQwen3.7 Plus—Not directly comparable
OpenHarmony Bench
MiniMax M348.4%SourceQwen3.7 Plus—Not directly comparable
LiveCodeBench (Vals)
MiniMax M382.2%SourceQwen3.7 Plus—Not directly comparable
SWE-bench (Vals)
MiniMax M375.0%SourceQwen3.7 Plus—Not directly comparable
SWE Multilingual
MiniMax M3—Qwen3.7 Plus75.8%SourceNot directly comparable
SciCode
MiniMax M3—Qwen3.7 Plus51.3%SourceNot directly comparable
LiveCodeBench
MiniMax M3—Qwen3.7 Plus89.6%SourceNot directly comparable
Reasoning
Knowledge
GPQA Diamond (Vals)
MiniMax M392.7%SourceQwen3.7 Plus—Not directly comparable
MMLU-Pro (Vals)
MiniMax M384.2%SourceQwen3.7 Plus—Not directly comparable
GPQA
MiniMax M3—Qwen3.7 Plus90.3%SourceNot directly comparable
GPQA-D
MiniMax M3—Qwen3.7 Plus90.3%SourceNot directly comparable
HLE
MiniMax M3—Qwen3.7 Plus34.7%SourceNot directly comparable
MMLU-Pro
MiniMax M3—Qwen3.7 Plus88.5%SourceNot directly comparable
MMLU-Redux
MiniMax M3—Qwen3.7 Plus94.5%SourceNot directly comparable
SuperGPQA
MiniMax M3—Qwen3.7 Plus71.4%SourceNot directly comparable
MMMLU
MiniMax M3—Qwen3.7 Plus89.0%SourceNot directly comparable
Math
Multilingual
MMLU-ProX
MiniMax M3—Qwen3.7 Plus85.4%SourceNot directly comparable
NOVA-63
MiniMax M3—Qwen3.7 Plus58.8%SourceNot directly comparable
INCLUDE
MiniMax M3—Qwen3.7 Plus83.0%SourceNot directly comparable
MAXIFE
MiniMax M3—Qwen3.7 Plus88.8%SourceNot directly comparable
PolyMath
MiniMax M3—Qwen3.7 Plus84.0%SourceNot directly comparable
Multimodal
OfficeQA Pro
MiniMax M345.1%SourceQwen3.7 Plus—Not directly comparable
OmniDocBench 1.5
MiniMax M391.6%SourceQwen3.7 Plus91.4%SourceMiniMax M3 leads this result
MMMU-Pro
MiniMax M378.1%SourceQwen3.7 Plus79%SourceQwen3.7 Plus leads this result
VideoMMMU
MiniMax M384.6%SourceQwen3.7 Plus85.4%SourceQwen3.7 Plus leads this result
Video-MME (with subtitle)
MiniMax M385.4%SourceQwen3.7 Plus88.0%SourceQwen3.7 Plus leads this result
MathVision
MiniMax M3—Qwen3.7 Plus90.3%SourceNot directly comparable
CharXiv
MiniMax M3—Qwen3.7 Plus85.9%SourceNot directly comparable
ERQA
MiniMax M3—Qwen3.7 Plus69.8%SourceNot directly comparable
MedXpertQA (MM)
MiniMax M3—Qwen3.7 Plus71.0%SourceNot directly comparable
ScreenSpot Pro
MiniMax M3—Qwen3.7 Plus79.0%SourceNot directly comparable
SimpleVQA
MiniMax M3—Qwen3.7 Plus81.7%SourceNot directly comparable
MMSearch-Plus
MiniMax M3—Qwen3.7 Plus41.4%SourceNot directly comparable
RealWorldQA
MiniMax M3—Qwen3.7 Plus86.9%SourceNot directly comparable
OCRBench V2
MiniMax M3—Qwen3.7 Plus70.7%SourceNot directly comparable
ODINW13
MiniMax M3—Qwen3.7 Plus51.1%SourceNot directly comparable
MLVU (M-Avg)
MiniMax M3—Qwen3.7 Plus87.4%SourceNot directly comparable
Instruction following
Frequently asked questions
Which is better, MiniMax M3 or Qwen3.7 Plus?
Qwen3.7 Plus has the higher public score estimate, 61.57 versus 61.51, 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 Qwen3.7 Plus?
Qwen3.7 Plus scores higher for coding on the public lane, 52.6 to 48.9. MiniMax M3 and Qwen3.7 Plus are 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 Qwen3.7 Plus?
MiniMax M3 leads the public agentic tasks lane, 42 to 37.3, with Supported evidence for both models, although the 90% intervals overlap.
Which costs less, MiniMax M3 or Qwen3.7 Plus?
Both models do not have comparable published API token rates, so this page does not name a universal price winner.
Which has the larger context window, MiniMax M3 or Qwen3.7 Plus?
Both models list the same context window, 1M.
Related comparisons
Last updated September 10, 2026
Watch MiniMax M3 vs Qwen3.7 Plus
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 2,000+ readers.