CodingActive
Trust Score:
90

SWE-bench Pro

Long-horizon repository tasks across public, held-out, and commercial codebases, with a continuing public model leaderboard.

Launched: Refresh: static
Status Assessment (active):

The benchmark still has a maintained public leaderboard and substantial score spread: the 2026-08-03 LLM Stats snapshot ranges from 0.584 to 0.800 across these 17 leading models, despite documented task-quality concerns.

Performance Timeline

Longitudinal progression of model scores against human baselines.

Performance & Historical Trajectory

Empirical score progression across model release dates and evaluation rounds.

Independent Vendor-Reported Human Baseline (74.0%)
ModelScoreDateSource TypeProvenance
GPT-5.6 Terra (LLM Stats snapshot)63.4%2024-08-01vendor-reportedSource ↗
GLM-5.1 (LLM Stats snapshot)58.4%2024-08-01vendor-reportedSource ↗
GPT-5.5 (LLM Stats snapshot)58.6%2024-08-01vendor-reportedSource ↗
Kimi K2.6 (LLM Stats snapshot)58.6%2024-08-01vendor-reportedSource ↗
Gemini 3.6 Flash (LLM Stats snapshot)58.7%2024-08-01vendor-reportedSource ↗
MiniMax M3 (LLM Stats snapshot)59%2024-08-01vendor-reportedSource ↗
Qwen3.7 Max (LLM Stats snapshot)60.6%2024-08-01vendor-reportedSource ↗
Muse Spark 1.1 (LLM Stats snapshot)61.5%2024-08-01vendor-reportedSource ↗
GLM-5.2 (LLM Stats snapshot)62.1%2024-08-01vendor-reportedSource ↗
GPT-5.6 Luna (LLM Stats snapshot)62.7%2024-08-01vendor-reportedSource ↗
Claude Sonnet 5 (LLM Stats snapshot)63.2%2024-08-01vendor-reportedSource ↗
Claude Opus 4.7 (LLM Stats snapshot)64.3%2024-08-01vendor-reportedSource ↗
GPT-5.6 Sol (LLM Stats snapshot)64.6%2024-08-01vendor-reportedSource ↗
Grok 4.5 (LLM Stats snapshot)64.7%2024-08-01vendor-reportedSource ↗
Claude Opus 4.8 (LLM Stats snapshot)69.2%2024-08-01vendor-reportedSource ↗
Claude Mythos Preview (LLM Stats snapshot)77.8%2024-08-01vendor-reportedSource ↗
Claude Fable 5 (LLM Stats snapshot)80%2024-08-01vendor-reportedSource ↗
Qwen3 32B + SWE-agent (public, launch revision)3.4%2024-06-01independentSource ↗
GPT-5 + SWE-agent (public, launch revision)23.3%2024-06-01independentSource ↗
Claude Opus 4.1 + SWE-agent (public, launch revision)22.7%2024-06-01independentSource ↗
Claude Sonnet 4 + SWE-agent (public, launch revision)17.6%2024-06-01independentSource ↗
GPT-4o + SWE-agent (public, launch revision)4.9%2024-06-01independentSource ↗

Human Baseline & Difficulty Horizon

Calibrated human reference points, specialist benchmarks, and ceiling thresholds.
Measured Human Score74.0%Domain Expert Baseline
Baseline Protocol & Interpretation

Professional software engineers completing repository-scale enterprise feature requests and bug fixes in 4 to 12 hours.

Metric & Scoring Methodology

Verification protocols, aggregation formulas, and specialized metric variants.
Primary Metric:resolved instances on the public set (%)
Scoring Engine:unit-tests

Dataset & Compute Cost

Evaluation volume, public availability, API pricing, and local hardware requirements.
Total Dataset Size731Annotated evaluation items
Public Test SetPublicOpenly mirrored on repositories
Access GatingOpen AccessUnrestricted download
Evaluation LicenseOpen AccessDataset usage and redistribution terms
Frontier API Compute Cost:

$5 – $20 USD for full benchmark evaluation run on frontier APIs.

Recommended Local GPU Setup:

1x NVIDIA RTX 4090 (24GB) or A100 (40GB/80GB) via vLLM / SGLang

Official Dataset & Benchmark Files:Download / View Dataset Repository ↗

How to Run & Reproduce

Standardized evaluation protocols, CLI commands, and reproducible runner templates.
Prompt Regimezero-shot
Reasoning Modedirect
Sampling Temp0
Pass@k Budgetk = 1
Tools & SandboxPure Text
Scoring Verifierunit-tests
Option AEleutherAI LM-Evaluation-Harness (Open-Weight Models)
lm_eval --model hf --model_args pretrained=<model_path> --tasks swe-bench-pro --batch_size auto
Option BOpenCompass Evaluation Framework
opencompass --datasets swe-bench-pro --models <model_config>
Python APIDeterministic Inference Loop Snippet
# Standard API Evaluation Loop
from openai import OpenAI

client = OpenAI()
response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": prompt}],
    temperature=0.0,
)
Standardized Reporting Requirement:

When publishing results for SWE-bench Pro, always report the exact prompt template, few-shot exemplar ordering, sampling temperature (temperature=0), maximum reasoning budget tokens, and the precise timestamped model snapshot ID.

Contamination & Memorization Analysis

Audit of pretraining exposure risks, memorization vectors, and refresh policies.
Overall Contamination Risk:MEDIUM
Refresh Cadence:

Static fixed snapshot

Test Set Exposure:

Public on web / HuggingFace