Step 5 Preview.
A model-level benchmark profile for Step 5 Preview, combining Arena category ranks, suite indexes, commercial price and speed data, SWE-bench status, and lab context.
Lab
Stepfun
Proprietary
Coverage
67%
6 of 9 tracked signals
Best Arena rank
#36
WebDev Arena
Best suite
#33
Coding / 38
Model profile
Benchmark signal map.
This view treats missing values as coverage gaps, then puts every available score, rank, cost, and speed signal into the same visual hierarchy.
Analyst view
Step 5 Preview is strongest on coding.
Its best composite result is rank #33 in the coding view, with a 38 coding index. The strongest raw Arena signal is WebDev Arena at rank #36 with 4,424 public votes.
Best current lane
Coding
#33 / 38 coding index
Strongest raw signal
Code
#36, 1565 score
Commercial posture
$1.43
1.39s first token / 91 tok/s
Coverage gaps
Profile completeness
6 of 9 signals matched
Arena categories
3/5
Code, Vision, Text
Commercial match
Live
Step 5 Preview
SWE-bench
Not listed
No confident software benchmark row
Raw identifier
Step 5 Preview
Performance cockpit
Public benchmark scores
Arena rows show score, public rank, confidence interval, vote count, and publish date. Missing rows stay visible as coverage gaps.
WebDev Arena
Arena category
1555-1575 CI
Vision Arena
Arena category
1232-1268 CI
Search Arena
Arena category
Not listed
No confident public Arena row for this category.
Coverage gap
Document Arena
Arena category
Not listed
No confident public Arena row for this category.
Coverage gap
SWE-bench Verified
Software issue resolution
Not listed
No confident SWE-bench row matched to this model.
Coverage gap
Cost and speed
Commercial profile
Price, latency, and throughput are only shown when a confident commercial source row is matched.
Blended price
$1.43
USD / 1M tokens
Input price
$1.00
USD / 1M input tokens
Output price
$2.70
USD / 1M output tokens
Latency
1.39s
Median time to first token
Output speed
91 tok/s
Median generation speed
Matched source row
Step 5 PreviewSuite position
Composite ranks by benchmark lens.
These views combine available public signals into purpose-specific indexes. A missing card means the model lacks enough source coverage for that lens.
Value
Quality, price, latency, and output speed
value index
Coding
Code Arena, SWE-bench, and output speed
coding index
Agents
Search, documents, code, and latency
agent proxy
Reasoning
Text, documents, vision, and mode signal
reasoning proxy
Long context
Document, search, and text benchmark proxy
context proxy