Chat turn cost
1K fresh input + 500 output tokens
LongCat-2.5
LongCat-2.5 has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-rates
LongCat-2.5 vs Step 5 Preview
Updated November 9, 2026. We do not rank this pair: at least one has no public score. Public scores include evidence status and uncertainty.
Share or export
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. 10 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.
- Shared results
- 10
- LongCat-2.5 only
- 12
- Step 5 Preview only
- 14
- Like-for-like categories
- 0 / 8
Estimated: Step 5 Preview. Conditional ranges do not establish rank confidence.How the comparison works
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.
Cache-heavy agent loop cost
200K cached + 20K fresh input + 10K output tokens
LongCat-2.5
LongCat-2.5 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
LongCat-2.5
LongCat-2.5 has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-ratesCoding work
Code generation, repair, and software-engineering tasks
Not enough matched evidence
LongCat-2.5 is not ranked on the public lane for coding, so no winner is named for coding.
Confidence: limitedAgentic work
Tool use, computer use, and multi-step task completion
Not enough matched evidence
LongCat-2.5 is not ranked on the public lane for agentic, so no winner is named for agentic.
Confidence: limitedLong 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.
—LongCat-2.549.1Step 5 Preview
Not comparable · BenchAlign v5.8
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.
Same basis rules as the category table below
What is actually comparable
Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.
12 LongCat-2.5 only
10 Shared
14 Step 5 Preview only
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.
AutomationBenchAgentic
Normalized gap 7.8LongCat-2.5:51.8%Step 5 Preview:44.0%BrowseCompAgentic
Normalized gap 3.2LongCat-2.5:91.9%Step 5 Preview:88.7%
Bars run 0–100 on each benchmark’s normalized display scale
Category results, on a stated basis
Each row shows the public-lane category score for both models: the BenchAlign v5.8 lane for agentic, coding, reasoning, multimodal, 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.
Agentic
Not comparable- LongCat-2.5
- Not ranked
- Step 5 Preview
- 54.9
- Estimated · #45/132
- Basis
- BenchAlign v5.8 lane · 11 vs 13 public rows
- Reading
- Not comparable
Coding
Not comparable- LongCat-2.5
- Not ranked
- Step 5 Preview
- 49.1
- Estimated · #50/158
- Basis
- BenchAlign v5.8 lane · 6 vs 6 public rows
- Reading
- Not comparable
Reasoning
Not comparable- LongCat-2.5
- Not ranked
- Step 5 Preview
- 52.5
- Estimated · #48/157
- Basis
- BenchAlign v5.8 lane · 0 vs 1 public rows
- Reading
- Not comparable
Multimodal
Not comparable- LongCat-2.5
- Not ranked
- Step 5 Preview
- 59.5
- Estimated · #38/67
- Basis
- BenchAlign v5.8 lane · 4 vs 2 public rows
- Reading
- Not comparable
Knowledge
Not comparable- LongCat-2.5
- Not ranked
- Step 5 Preview
- 60.2
- Supported · #45/179
- Basis
- BenchAlign v5.8 lane · 1 vs 2 public rows
- Reading
- Not comparable
Multilingual
Not comparable- LongCat-2.5
- Not ranked
- Step 5 Preview
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Instruction following
Not comparable- LongCat-2.5
- Not ranked
- Step 5 Preview
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Math
Not comparable- LongCat-2.5
- Not ranked
- Step 5 Preview
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
| Category | LongCat-2.5 | Step 5 Preview | Basis | Reading |
|---|---|---|---|---|
| Agentic | Not ranked | 54.9Estimated · #45/132 | Not comparableBenchAlign v5.8 lane · 11 vs 13 public rows | Not comparable |
| Coding | Not ranked | 49.1Estimated · #50/158 | Not comparableBenchAlign v5.8 lane · 6 vs 6 public rows | Not comparable |
| Reasoning | Not ranked | 52.5Estimated · #48/157 | Not comparableBenchAlign v5.8 lane · 0 vs 1 public rows | Not comparable |
| Multimodal | Not ranked | 59.5Estimated · #38/67 | Not comparableBenchAlign v5.8 lane · 4 vs 2 public rows | Not comparable |
| Knowledge | Not ranked | 60.2Supported · #45/179 | Not comparableBenchAlign v5.8 lane · 1 vs 2 public rows | Not comparable |
| Multilingual | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Instruction following | Not ranked | 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, reasoning, multimodal, and knowledge denominators (BenchAlign v5.8) differ from the provisional-lane categories. Unranked scores sit on the provisional lane’s scale but never name a winner. Like-for-like rows within 0.5 points read as a practical tie.
Supported evidence per lane · bars run 0–100Methodology
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
- LongCat-2.5
- $0.0009
- Fits in one request
- Step 5 Preview
- $0.00235
- Fits in one request
LongCat-2.5 has the lower modeled cost
Costs use the listed standard API rates.
Repository review
50K fresh input + 3K output tokens
- LongCat-2.5
- $0.0186
- Fits in one request
- Step 5 Preview
- $0.0581
- Fits in one request
LongCat-2.5 has the lower modeled cost
Costs use the listed standard API rates.
Cache-heavy agent loop
200K cached + 20K fresh input + 10K output tokens
- LongCat-2.5
- $0.0192
- Fits in one request
- Step 5 Preview
- $0.057
- Fits in one request
LongCat-2.5 has the lower modeled cost
Costs use the listed standard API rates.
Cached input falls back to the list input rate only where a cached rate is unpublished
Specification differences
Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.
SpecificationLongCat-2.5Step 5 Preview
Context window
Maximum documented context; output-token limits may be lower.
LongCat-2.5
Step 5 Preview
API model ID
LongCat-2.5
LongCat-2.5
LongCat API Platform pricingStep 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.
LongCat-2.5
$0.006 per 1M cached input tokens
LongCat API Platform pricingStep 5 Preview
$0.05 per 1M cached input tokens
StepFun pricing and rate limitsDocumented inputs
LongCat-2.5
Not sourced
Step 5 Preview
text, image, video
StepFun Step 5 Preview model documentationDocumented outputs
LongCat-2.5
Not sourced
Step 5 Preview
Provider availability
LongCat-2.5
Not sourced
Step 5 Preview
Preview · StepFun Open Platform API, Hugging Face, self-hosted inference
StepFun Step 5 Preview model documentationReasoning profile
LongCat-2.5
Hybrid
Step 5 Preview
Reasoning
Weight access
LongCat-2.5
Open Weight
Step 5 Preview
Open Weight
License
LongCat-2.5
Open Weight
Step 5 Preview
Open Weight
Release date
LongCat-2.5
2026-11-02
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: $0.0192 vs $0.057.
- Context tradeoff
- Both models list 1M.
Run the same representative tasks against both endpoints before changing production traffic.
Questions
Which is better, LongCat-2.5 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, LongCat-2.5 or Step 5 Preview?
LongCat-2.5 is not ranked on the public lane for coding, so no winner is named for coding.
Which is better for agentic tasks, LongCat-2.5 or Step 5 Preview?
LongCat-2.5 is not ranked on the public lane for agentic tasks, so no winner is named for agentic tasks.
Which costs less, LongCat-2.5 or Step 5 Preview?
For the stated presets, chat costs $0.0009 on LongCat-2.5 and $0.00235 on Step 5 Preview; repository review costs $0.0186 and $0.0581; the cache-heavy agent loop costs $0.0192 and $0.057. Costs use the listed standard API rates.
Which has the larger context window, LongCat-2.5 or Step 5 Preview?
Both models list the same context window, 1M.
Self-host vs API cost
Estimates at 50,000 req/day · 1000 tokens/req average.
LongCat-2.5
API / mo$1,125
Self-host / mo$57,606
Break-even3.4B/day
Step 5 Preview
API / mo$2,775
Self-host / mo$27,962
Break-even617M/day
Benchmark evidence
The full public result ledger is available for audit without forcing a wide desktop table onto a phone.
Browse raw public benchmark evidence36 rows
Agentic
Terminal-Bench 2.1
LongCat-2.588.4%SourceStep 5 Preview85.0%SourceLongCat-2.5 leads this result
Terminal-Bench 4.0
LongCat-2.538.50%SourceStep 5 Preview33.30%SourceLongCat-2.5 leads this result
BrowseComp
LongCat-2.591.9%SourceStep 5 Preview88.7%SourceLongCat-2.5 leads this result
AutomationBench
LongCat-2.551.8%SourceStep 5 Preview44.0%SourceLongCat-2.5 leads this result
HLE w/ tools
LongCat-2.563.1%SourceStep 5 Preview—Not directly comparable
Agents' Last Exam
LongCat-2.538.5%SourceStep 5 Preview29.5%SourceLongCat-2.5 leads this result
MCP Atlas
LongCat-2.583.1%SourceStep 5 Preview85.6%SourceStep 5 Preview leads this result
JobBench
LongCat-2.565.4%SourceStep 5 Preview59.0%SourceLongCat-2.5 leads this result
skillsBench
LongCat-2.563.3%SourceStep 5 Preview—Not directly comparable
SpreadsheetBench 2
LongCat-2.534.9%SourceStep 5 Preview—Not directly comparable
Claw-Eval
LongCat-2.578.3%SourceStep 5 Preview—Not directly comparable
CyberGym
LongCat-2.5—Step 5 Preview84.7%SourceNot directly comparable
Toolathlon-Verified
LongCat-2.5—Step 5 Preview74.1%SourceNot directly comparable
APEX-Agents
LongCat-2.5—Step 5 Preview37.8%SourceNot directly comparable
DRACO
LongCat-2.5—Step 5 Preview83.3%SourceNot directly comparable
Hermes Index
LongCat-2.5—Step 5 Preview32.8%SourceNot directly comparable
Hermes Bench
LongCat-2.5—Step 5 Preview69.2%SourceNot directly comparable
Coding
Terminal-Bench 2.1
LongCat-2.588.4%SourceStep 5 Preview85.0%SourceLongCat-2.5 leads this result
ProgramBench
LongCat-2.525.5%SourceStep 5 Preview80.5%SourceStep 5 Preview leads this result
DeepSWE
LongCat-2.564.6%SourceStep 5 Preview67.7%SourceStep 5 Preview leads this result
PaperBench
LongCat-2.582.1%SourceStep 5 Preview—Not directly comparable
SWE-Atlas Codebase QnA
LongCat-2.558.3%SourceStep 5 Preview—Not directly comparable
NL2Repo
LongCat-2.560.5%SourceStep 5 Preview—Not directly comparable
SciCode
LongCat-2.5—Step 5 Preview58.9%SourceNot directly comparable
sweMarathon
LongCat-2.5—Step 5 Preview72.7%SourceNot directly comparable
MLS-Bench Lite
LongCat-2.5—Step 5 Preview40.5%SourceNot directly comparable
Reasoning
CritPt
LongCat-2.5—Step 5 Preview20.9%SourceNot directly comparable
Multimodal
Vision2Web
LongCat-2.578.6%SourceStep 5 Preview—Not directly comparable
CharXiv
LongCat-2.590.2%SourceStep 5 Preview—Not directly comparable
MathVision w/ Python
LongCat-2.596.5%SourceStep 5 Preview—Not directly comparable
Video-MME
LongCat-2.585.2%SourceStep 5 Preview—Not directly comparable
OfficeQA Pro
LongCat-2.5—Step 5 Preview60.3%SourceNot directly comparable
MMMU-Pro
LongCat-2.5—Step 5 Preview76%SourceNot directly comparable
Knowledge
36 public results · 10 shared
Watch LongCat-2.5 vs Step 5 Preview
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 6,100+ subscribers.
Last updated November 9, 2026