Model results

DeepSWE

Choose a shared evaluation protocol to compare model scores. Versions, sources and configurations stay separate.

Atlas snapshot captured 10/19/2026, 10:20:28 AM UTC. This is an import timestamp, not the evaluation date. Coverage may be incomplete.

Reviewed recommendation and limits

Use as supplementary evidence for original repository engineering tasks, with the grading audit beside the result.

DeepSWE v1.1. Grade committed patches in a separate verifier container. Preserve agent, effort, repetitions and completed-trial denominator; keep v1 and later revisions separate.. Imported groups below may use different protocols; this review does not certify each run. Samaritan Research reports false-negative grading defects in at least 23 of 113 v1.1 tasks. A subsequent repair has not been verified in this review, so affected results need caution.

Publisher results Publisher v1.1 methods First-hand grading audit

← All evaluation groups

DeepSWE v1.1

Source: DeepSWE via Epoch AI
Version: v1.1
Measurement date: Not reported
Metric: % · Higher is better
Evaluation scope: Harness mini-swe-agent · result date not supplied

Positions are withheld because at least one result lacks explicit independent provenance or a documented evaluation scope. Missing metadata is shown as unreported; matching imported metadata is not independent verification of the protocol.

Muse Spark 1.1
meta/muse-spark-1.1-20260709
53.3%
View modelEvidence aggregated
Kimi K2.7 Code
moonshotai/kimi-k2.7-code-20260612
30.5%
View modelEvidence aggregated
Gemini 3.1 Pro Preview
google/gemini-3.1-pro-preview-20260219
11.7%
View modelEvidence aggregated