Atlas

Benchmarks

Longtail-SWE (Pass@3)

Longtail Swe

The share of Longtail-SWE's 202 issue-resolution tasks solved in at least one of three attempts, in the authors' SWE-agent-like scaffold.

Models
7
Top score
89.1
Leader
Claude Opus 5.5
Updated
Nov 6, 2026
Resolution
—
Saturation
—

No signal or difficulty estimate. Too few models have been measured on this benchmark to estimate its difficulty.

Organizations
7 dated models · 3 frontier modelsClosed weightsOpen weights
Data: Nebiusatlas.village.dev
#ModelSettingScoreSource
1Claude Opus 5.5high · SWE-rebench swelike scaffold; 1M context
89.1
2GPT-5.6 Solhigh · SWE-rebench swelike scaffold; 1M context
87.1
3GPT-5.6 Lunax-high · SWE-rebench swelike scaffold; 1M context
84.2
4GPT-6.1 Solhigh · SWE-rebench swelike scaffold; 1M context
82.2
5Kimi K3max · SWE-rebench swelike scaffold; 262K context
76.2
6GLM-5.3 Flashnot reported · SWE-rebench swelike scaffold; 262K context
69.8
7Qwen3.6 27Bthinking · SWE-rebench swelike scaffold; 262K context
23.8
Loading Atlas data…