AI IQ
Subscribe

Backend Engineering

A bird's-eye view of the dimension: the main IQ chart, the effective-cost frontier, and the benchmark charts that feed it.

Backend Engineering IQ
Backend Engineering IQ
Each model's Backend Engineering IQ plotted on a standard normal IQ distribution
Controls:

How to read this chart

Models are placed on a standard IQ-style distribution using Backend Engineering IQ. Higher scores appear farther to the right.

Benchmarks in this dimension

SWE-Bench VerifiedLiveCodeBenchSWE-Bench ProAPEX-SWEDeepSWEFrontierSWEFrontierCode DiamondSWE-rebenchDeepSWE v1.1SWE Marathon
Backend Engineering vs Effective Cost
Backend Engineering vs Effective Cost
X = effective cost (log). Y = Backend Engineering IQ. Color = provider.
Controls:
Production Eng.1:1Cost

How to read this chart

Each point is a public model. The chart compares Backend Engineering IQ against Effective Cost (per 1M I/O Tokens), with color showing the model provider.

Data sources
Artificial AnalysisARC PrizeVals.aiCognition FrontierCodeMercor APEX-SWELLM StatsScale LabsScale Labs SWE-Bench Pro Public DatasetDeepSWE v1.1DeepSWESWE-rebenchSWE Marathon
Benchmarks In This Dimension
SWE-Bench Verified Benchmark Scores
Each model's SWE-Bench Verified score. Color = provider.
Controls:

How to read this chart

Bars rank models by the source-backed benchmark value used for this chart. Longer bars indicate higher published scores.

Data sources
Vals.aiLLM Stats
LiveCodeBench Benchmark Scores
Each model's LiveCodeBench score. Color = provider.
Controls:

How to read this chart

Bars rank models by the source-backed benchmark value used for this chart. Longer bars indicate higher published scores.

Data sources
Vals.ai
SWE-Bench Pro Benchmark Scores
Each model's SWE-Bench Pro percentage score. Color = provider.
Controls:

How to read this chart

Bars rank models by the source-backed benchmark value used for this chart. Longer bars indicate higher published scores.

Data sources
Scale LabsScale Labs SWE-Bench Pro Public Dataset
APEX-SWE Scores
Mercor APEX-SWE overall Pass@1 scores for economically valuable software-engineering work. Color = provider.
Controls:

How to read this chart

Bars rank models by the source-backed benchmark value used for this chart. Longer bars indicate higher published scores.

Data sources
Mercor APEX-SWE
DeepSWE Benchmark Scores
Each model's DeepSWE Pass@1 score. Color = provider.
Controls:

How to read this chart

Bars rank models by the source-backed benchmark value used for this chart. Longer bars indicate higher published scores.

Data sources
DeepSWE
FrontierSWE Scores
FrontierSWE leaderboard tracking for software-engineering skill at the edge of human ability. Color = provider.
Controls:

How to read this chart

Bars rank models by the source-backed benchmark value used for this chart. Longer bars indicate higher published scores.

Data sources
FrontierSWE
FrontierCode Diamond Scores
Each model's FrontierCode Diamond score for high-quality production-code tasks. Color = provider.
Controls:

How to read this chart

Bars rank models by the source-backed benchmark value used for this chart. Longer bars indicate higher published scores.

Data sources
Cognition FrontierCode
SWE-rebench Benchmark Scores
Each model's SWE-rebench resolved rate. Color = provider.
Controls:

How to read this chart

Bars rank models by the source-backed benchmark value used for this chart. Longer bars indicate higher published scores.

Data sources
SWE-rebench
DeepSWE v1.1 Benchmark Scores
Each model's DeepSWE v1.1 Pass@1 score. Color = provider.
Controls:

How to read this chart

Bars rank models by the source-backed benchmark value used for this chart. Longer bars indicate higher published scores.

Data sources
DeepSWE v1.1
SWE Marathon Scores
Each model's SWE Marathon aggregate Resolution Rate (Pass@1). Color = provider.
Controls:

How to read this chart

Bars rank models by the source-backed benchmark value used for this chart. Longer bars indicate higher published scores.

Data sources
SWE Marathon