Atlas

Benchmarks

CrossSearch (pass@1)

CrossSearch's pass@1: the share of all 300 attempts (three on each of 100 tasks) that earn full credit, meaning every rubric item for the task passes under the GPT-6.1 Sol judge. Unlike the pass^3 headline, it credits tasks a model solves only sometimes.

Models
13
Top score
72.3
Leader
Gemini 4 Argon
Updated
Nov 12, 2026
Resolution
—
Saturation
—
Signal
91stpercentile
195 benchmarks from 0.8 to 3.5224 benchmarks from 3.5 to 6.1133 benchmarks from 6.1 to 8.898 benchmarks from 8.8 to 11.566 benchmarks from 11.5 to 14.126 benchmarks from 14.1 to 16.815 benchmarks from 16.8 to 19.59 benchmarks from 19.5 to 22.18 benchmarks from 22.1 to 24.84 benchmarks from 24.8 to 27.55 benchmarks from 27.5 to 30.12 benchmarks from 30.1 to 32.84 benchmarks from 32.8 to 35.45 benchmarks from 35.4 to 38.11 benchmark from 38.1 to 40.82 benchmarks from 43.4 to 46.11 benchmark from 62.1 to 64.8

73rd strongest of 798

Signal-to-noise +14.98 for the CrossSearch (pass^3) family

Difficulty
Not estimated

Another benchmark represents this family in the fit.

Organizations
13 dated models · 4 frontier modelsClosed weightsOpen weights
Data: AfterQueryatlas.village.dev

CrossSearch (pass@1) leaderboard

#ModelSettingScoreSource
1Gemini 4 Argonhigh · AfterQuery CrossSearch harness: Terminus 2 on Harbor, one shell with live internet, no search tool and a view_image command; three attempts per task; 65,536-token response limit
72.3
2GPT-6 Astrax-high · AfterQuery CrossSearch harness: Terminus 2 on Harbor, one shell with live internet, no search tool and a view_image command; three attempts per task; 128,000-token response limit
71.3
3Claude Opus 5.5x-high · AfterQuery CrossSearch harness: Terminus 2 on Harbor, one shell with live internet, no search tool and a view_image command; three attempts per task; 128,000-token response limit
70.3
4GPT-6.1 Solx-high · AfterQuery CrossSearch harness: Terminus 2 on Harbor, one shell with live internet, no search tool and a view_image command; three attempts per task; 128,000-token response limit
68.0
5Claude Fable 5.2high · AfterQuery CrossSearch harness: Terminus 2 on Harbor, one shell with live internet, no search tool and a view_image command; three attempts per task; 32,768-token response limit
66.7
6GPT-5.6 Solx-high · AfterQuery CrossSearch harness: Terminus 2 on Harbor, one shell with live internet, no search tool and a view_image command; three attempts per task; 128,000-token response limit
59.0
7Muse Wonderhigh · AfterQuery CrossSearch harness: Terminus 2 on Harbor, one shell with live internet, no search tool and a view_image command; three attempts per task; 32,768-token response limit
58.3
8Claude Opus 5x-high · AfterQuery CrossSearch harness: Terminus 2 on Harbor, one shell with live internet, no search tool and a view_image command; three attempts per task; 128,000-token response limit
54.3
9Grok 4.6x-high · AfterQuery CrossSearch harness: Terminus 2 on Harbor, one shell with live internet, no search tool and a view_image command; three attempts per task; 450,000-token response limit
53.3
10Gemini 3.8 Flashhigh · AfterQuery CrossSearch harness: Terminus 2 on Harbor, one shell with live internet, no search tool and a view_image command; three attempts per task; 65,536-token response limit
45.0
11Inkling 2 Smallthinking · AfterQuery CrossSearch harness: Terminus 2 on Harbor, one shell with live internet, no search tool and a view_image command; three attempts per task; 262,144-token response limit; Tinker's 256K-context listing, asked for effort "high", whose numeric setting is undocumented
44.0
12Kimi K3high · AfterQuery CrossSearch harness: Terminus 2 on Harbor, one shell with live internet, no search tool and a view_image command; three attempts per task; 943,718-token response limit
32.3
13Sonnet 5x-high · AfterQuery CrossSearch harness: Terminus 2 on Harbor, one shell with live internet, no search tool and a view_image command; three attempts per task; 128,000-token response limit
27.0

Related benchmarks

Most correlated

Benchmarks outside its family whose model rankings track it most reliably (Spearman ρ over at least 10 shared models).

Loading Atlas data…