Agentic
Not comparable- MiniMax M2.7
- 41.2
- Estimated · #113/154
- Step 5 Preview
- Not ranked
- Basis
- BenchAlign lane · 7 vs 11 public rows
- Reading
- Not comparable
MiniMax M2.7 vs Step 5 Preview
Decision reading
At least one model is not scored in the current public ranking lane, so the page does not name an overall quality winner.
1 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.
Updated September 21, 2026. We do not rank this pair: at least one has no public score. 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.
Long documents
Prompts that approach the documented context limit
Step 5 Preview
Step 5 Preview has the larger documented context window.
Confidence: documented
Chat turn cost
1K fresh input + 500 output tokens
MiniMax M2.7
MiniMax M2.7 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 M2.7
MiniMax M2.7 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
Step 5 Preview is not ranked on the public lane for coding, so no winner is named for coding.
Confidence: limited
Agentic work
Tool use, computer use, and multi-step task completion
Not enough matched evidence
Step 5 Preview is not ranked on the public lane for agentic, so no winner is named for agentic.
Confidence: limited
Cache-heavy agent loop cost
200K cached + 20K fresh input + 10K output tokens
Not enough matched evidence
The page does not recommend a cost winner because at least one model cannot fit the stated workload in one request. MiniMax M2.7 does not fit this workload in one request. MiniMax M2.7 has no published cached-input rate, so cached tokens use its listed input rate.
Confidence: rate-fallback
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.
48.8MiniMax M2.7—Step 5 Preview
Not comparable · BenchAlign
The coding row is not comparable on the public lane: at least one model is not measured or not ranked there.
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.
22 MiniMax M2.7 only
1 Shared
21 Step 5 Preview only
- Shared results
- 1
- MiniMax M2.7 only
- 22
- Step 5 Preview only
- 21
- Like-for-like categories
- 0 / 8
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
Not comparable- MiniMax M2.7
- 48.8
- Estimated · #72/156
- Step 5 Preview
- Not ranked
- Basis
- BenchAlign lane · 11 vs 6 public rows
- Reading
- Not comparable
Reasoning
Not comparable- MiniMax M2.7
- 74.8
- Unranked · 2 rankable rows
- Step 5 Preview
- 78.5
- Unranked · 2 rankable rows
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Multimodal
Not comparable- MiniMax M2.7
- Not ranked
- Step 5 Preview
- 61.2
- #28/49
- Basis
- Provisional lane · 0 vs 2 weighted rows
- Reading
- Not comparable
Knowledge
Not comparable- MiniMax M2.7
- 49.1
- Supported · #92/186
- Step 5 Preview
- Not ranked
- Basis
- BenchAlign lane · 4 vs 2 public rows
- Reading
- Not comparable
Multilingual
Not comparable- MiniMax M2.7
- Not ranked
- Step 5 Preview
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Instruction following
Not comparable- MiniMax M2.7
- 91.6
- #10/124
- Step 5 Preview
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Math
Not comparable- MiniMax M2.7
- Not ranked
- Step 5 Preview
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
| Category | MiniMax M2.7 | Step 5 Preview | Basis | Reading |
|---|---|---|---|---|
| Agentic | 41.2Estimated · #113/154 | Not ranked | Not comparableBenchAlign lane · 7 vs 11 public rows | Not comparable |
| Coding | 48.8Estimated · #72/156 | Not ranked | Not comparableBenchAlign lane · 11 vs 6 public rows | Not comparable |
| Reasoning | 74.8Unranked · 2 rankable rows | 78.5Unranked · 2 rankable rows | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Multimodal | Not ranked | 61.2#28/49 | Not comparableProvisional lane · 0 vs 2 weighted rows | Not comparable |
| Knowledge | 49.1Supported · #92/186 | Not ranked | Not comparableBenchAlign lane · 4 vs 2 public rows | Not comparable |
| Multilingual | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Instruction following | 91.6#10/124 | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Math | Not ranked | 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.
A shared-evidence shape is not available.
BenchLM does not draw a radar or infer missing axes when the matched evidence is too sparse.
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 M2.7
- $0.0009
- Fits in one request
- Step 5 Preview
- $0.00235
- Fits in one request
MiniMax M2.7 has the lower modeled cost
Costs use the listed standard API rates.
Repository review
50K fresh input + 3K output tokens
- MiniMax M2.7
- $0.0186
- Fits in one request
- Step 5 Preview
- $0.0581
- Fits in one request
MiniMax M2.7 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 M2.7
- $0.078
- Does not fit in one request
- Cached input priced at the published list-input rate
- Step 5 Preview
- $0.057
- Fits in one request
MiniMax M2.7 does not fit this workload in one request. MiniMax M2.7 has no published cached-input rate, so cached tokens use its listed input rate.
Specification differences
Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.
SpecificationMiniMax M2.7Step 5 Preview
Context window
Maximum documented context; output-token limits may be lower.
MiniMax M2.7
200K
Step 5 Preview
API model ID
MiniMax M2.7
Not sourced
Step 5 Preview
step-5-preview
StepFun Step 5 Preview model documentationCached-input rate
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
MiniMax M2.7
Not published
Step 5 Preview
$0.05 per 1M cached input tokens
StepFun pricing and rate limitsDocumented inputs
MiniMax M2.7
Not sourced
Step 5 Preview
text, image, video
StepFun Step 5 Preview model documentationDocumented outputs
MiniMax M2.7
Not sourced
Step 5 Preview
Provider availability
MiniMax M2.7
Not sourced
Step 5 Preview
Preview · StepFun Open Platform API
StepFun Step 5 Preview model documentationReasoning profile
MiniMax M2.7
Non-Reasoning
Step 5 Preview
Reasoning
Weight access
MiniMax M2.7
Open Weight
Step 5 Preview
Pending
License
MiniMax M2.7
Open Weight
Step 5 Preview
Pending
Release date
MiniMax M2.7
2026-03-18
Step 5 Preview
2026-09-20
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
- At least one model is not scored in the current public ranking lane, so the page does not name an overall quality winner.
- Workload cost
- Repository review: $0.0186 vs $0.0581. Cache-heavy agent loop: $0.078 vs $0.057.
- Context tradeoff
- Step 5 Preview has the larger documented window (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 evidence44 rows
Agentic
Terminal-Bench 2.0
MiniMax M2.757%SourceStep 5 Preview—Not directly comparable
Toolathlon
MiniMax M2.746.3%SourceStep 5 Preview—Not directly comparable
MLE-Bench Lite
MiniMax M2.766.6%SourceStep 5 Preview—Not directly comparable
MM-ClawBench
MiniMax M2.762.7%SourceStep 5 Preview—Not directly comparable
Claw-Eval
MiniMax M2.748.7%SourceStep 5 Preview—Not directly comparable
Gert Labs
MiniMax M2.740.40%SourceStep 5 Preview—Not directly comparable
Terminal-Bench 2.1 (Vals)
MiniMax M2.748.7%SourceStep 5 Preview—Not directly comparable
Terminal-Bench 2.1
MiniMax M2.7—Step 5 Preview85.0%SourceNot directly comparable
Terminal-Bench 4.0
MiniMax M2.7—Step 5 Preview33.30%SourceNot directly comparable
CyberGym
MiniMax M2.7—Step 5 Preview84.7%SourceNot directly comparable
AutomationBench
MiniMax M2.7—Step 5 Preview44.0%SourceNot directly comparable
Toolathlon-Verified
MiniMax M2.7—Step 5 Preview74.1%SourceNot directly comparable
MCP Atlas
MiniMax M2.7—Step 5 Preview85.6%SourceNot directly comparable
JobBench
MiniMax M2.7—Step 5 Preview59.0%SourceNot directly comparable
APEX-Agents
MiniMax M2.7—Step 5 Preview37.8%SourceNot directly comparable
DRACO
MiniMax M2.7—Step 5 Preview83.3%SourceNot directly comparable
BrowseComp
MiniMax M2.7—Step 5 Preview88.7%SourceNot directly comparable
Agents' Last Exam
MiniMax M2.7—Step 5 Preview29.5%SourceNot directly comparable
Coding
SWE-bench Verified*
MiniMax M2.775.4%SourceStep 5 Preview—Not directly comparable
SWE-bench Pro
MiniMax M2.756.2%SourceStep 5 Preview—Not directly comparable
SWE-Rebench
MiniMax M2.751.9%SourceStep 5 Preview—Not directly comparable
SWE Multilingual
MiniMax M2.776.5%SourceStep 5 Preview—Not directly comparable
Multi-SWE Bench
MiniMax M2.752.7%SourceStep 5 Preview—Not directly comparable
VIBE-Pro
MiniMax M2.755.6%SourceStep 5 Preview—Not directly comparable
NL2Repo
MiniMax M2.739.8%SourceStep 5 Preview—Not directly comparable
Vibe Code Bench
MiniMax M2.727.04%SourceStep 5 Preview—Not directly comparable
React Native Evals
MiniMax M2.771.4%SourceStep 5 Preview—Not directly comparable
LiveCodeBench (Vals)
MiniMax M2.779.9%SourceStep 5 Preview—Not directly comparable
SWE-bench (Vals)
MiniMax M2.773.8%SourceStep 5 Preview—Not directly comparable
DeepSWE
MiniMax M2.7—Step 5 Preview67.7%SourceNot directly comparable
Terminal-Bench 2.1
MiniMax M2.7—Step 5 Preview85.0%SourceNot directly comparable
SciCode
MiniMax M2.7—Step 5 Preview58.9%SourceNot directly comparable
ProgramBench
MiniMax M2.7—Step 5 Preview80.5%SourceNot directly comparable
sweMarathon
MiniMax M2.7—Step 5 Preview72.7%SourceNot directly comparable
MLS-Bench Lite
MiniMax M2.7—Step 5 Preview40.5%SourceNot directly comparable
Reasoning
CritPt
MiniMax M2.7—Step 5 Preview20.9%SourceNot directly comparable
Multimodal
Knowledge
GPQA-D
MiniMax M2.787.0%SourceStep 5 Preview93.5%SourceStep 5 Preview leads this result
MMLU-Pro (Arcee)
MiniMax M2.780.8%SourceStep 5 Preview—Not directly comparable
GPQA Diamond (Vals)
MiniMax M2.786.6%SourceStep 5 Preview—Not directly comparable
MMLU-Pro (Vals)
MiniMax M2.780.4%SourceStep 5 Preview—Not directly comparable
HLE
MiniMax M2.7—Step 5 Preview46.5%SourceNot directly comparable
Math
AIME25 (Arcee)
MiniMax M2.780.0%SourceStep 5 Preview—Not directly comparable
Questions
Which is better, MiniMax M2.7 or Step 5 Preview?
At least one model is not scored in the current public ranking lane, so the page does not name an overall quality winner. The page therefore keeps the decision tied to the specific documented workload.
Which is better for coding, MiniMax M2.7 or Step 5 Preview?
Step 5 Preview is not ranked on the public lane for coding, so no winner is named for coding.
Which is better for agentic tasks, MiniMax M2.7 or Step 5 Preview?
Step 5 Preview is not ranked on the public lane for agentic tasks, so no winner is named for agentic tasks.
Which costs less, MiniMax M2.7 or Step 5 Preview?
For the stated presets, chat costs $0.0009 on MiniMax M2.7 and $0.00235 on Step 5 Preview; repository review costs $0.0186 and $0.0581; the cache-heavy agent loop costs $0.078 and $0.057. MiniMax M2.7 does not fit this workload in one request. MiniMax M2.7 has no published cached-input rate, so cached tokens use its listed input rate.
Which has the larger context window, MiniMax M2.7 or Step 5 Preview?
Step 5 Preview has the larger documented context window: 1M, compared with 200K.
Related comparisons
Last updated September 21, 2026
Watch MiniMax M2.7 vs Step 5 Preview
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 2,000+ readers.