Rank #11
Model result · rank #11
Step 3.7 Flash
OpenRouter · stepfun/step-3.7-flash · extra-high reasoning · Full + SWE + Hard measured. Public result card with the model’s overall score, lane measurements, runtime/cost telemetry, and ranking formula.
Full rank #3
SWE rank #7
Hard rank #11
100.0% reliability
All-around publication view
Score68.06
Formulamean(Full, SWE, Hard Intelligence)
BasisOpenRouter · stepfun/step-3.7-flash · extra-high reasoning · Full + SWE + Hard measured
The overall score averages the measured major lanes while keeping each source measurement visible.
Full / Agentic benchmark
Final89.79
Capability89.97
Agentic84.11
Pass rate88.4%
Prompts43
This lane captures instruction following, structured behavior, tool discipline, and general agentic reliability.
Software engineering MVP
SWE score80.39
Focused final71.22
Capability80.39
Daily driver67.40
Prompts24
This lane is closer to implementation usefulness: source handling, architecture cleanliness, and deliverable quality.
Hard Intelligence diagnostic
Hard score33.99
Active inquiry17.50
Online adaptation80.00
Self-repair29.46
Authority integrity9.00
Hard Intelligence measures active inquiry, online adaptation, evidence-driven self-repair, and authority/salience integrity.
Runtime economics
Full cost$0.119
SWE cost$0.260
Hard cost$0.116
Full avg seconds7.20
Hard records8
Decode104.69
Cost, time, and runtime basis are telemetry. They explain tradeoffs; they do not secretly overwrite the capability scores.
Interpretation
Why this result lands here.
The model is stronger in the Full/Agentic lane than in the SWE lane; the overall score is therefore shown with both component lanes visible. Hard Intelligence score is 33.99 and contributes to the overall score alongside Full/Agentic and SWE. Hard Intelligence is published alongside Full/Agentic and SWE for the current ranking.