MINES BENCHMARK · REASONING MEDIUM

GPT-6 Astra

Openai · released 2026-09-03 · route gpt-6-astra · Model profile · Chess results · Sudoku results · FreeCell results · Nations Royale results · Hustle Sim results · Elemental Armory results · Micro Mouse results

Results

Score
2595%: 0 to 52 · 4 of 24 base boards so far
Boards won
1 of 4
Safe cells cleared beyond the opening
58.7%
Won / hit a mine / unfinished / unreadable
1 / 3 / 0 / 0
Ended for changing nothing five turns running
0
Reached the turn cap
0
Flags left on safe cells
5 in 3 boards
Illegal actions
0 of 470 (0.0%)
Wasted actions (already open or flagged)
44 of 470 (9.4%)
Unreadable replies
0
Turns per board
7.2
Requests per board
7.2
API-equivalent cost per win
$3.62
Tokens per board (in / out)
152,244 / 8,541
Reasoning tokens per board
4,924
API-equivalent cost per board
$0.904
Seconds per board
340.6
Replies cut off at the length limit
0.0%
Aborted on our side (not counted)
0

By tier

TierBoard, minesWonClearedTurnsOutput tokensSeconds
Beginner9×9, 100 of 168%3.01,59051.9
Intermediate16×16, 400 of 163%5.07,026504.5
Expert30×16, 991 of 1100%16.020,795672.6
Master30×24, 2000 of 14%5.04,753133.2

Cells by the technique that proves them

Each safe cell covered at the start is tagged with the simplest deduction that first proves it safe, given everything a perfect player could open before it.

TechniqueCellsOpened
Cascade from an empty cell45860.0%
One number56939.2%
Two numbers together6833.8%
The mine count4100.0%

Every board