CHESS BENCHMARK · LOCAL, NO THINKING

Qwen3.8 27B

Alibaba · open weights · released 2026-08-14 · route qwen3.8:27b · Model profile

Opening gap: +97 (95%: -845 to +1,039). The interval includes zero, so we can't yet say whether it performs differently across the variants.

Standard

Rating
80995%: 110 to 1,509
Record
0 won, 0 drawn, 1 lost of 1
Accuracy (open positions)
73.1
Accuracy (all moves)
77.3
Average centipawn loss
162
Blunders / mistakes / inaccuracies per 100 moves
16.7 / 16.7 / 0.0
Matched Stockfish's first choice
42%
In Stockfish's top three
67%
Legal on the first try
92.3%
Illegal / invalid
3 of 15 answers
Forfeits
1
Adjudicated games
0
API-equivalent cost per game
n/a
API-equivalent cost per move
n/a
Seconds per move
735.5
Model minutes per game
159.4
Tokens per move (in / out)
292 / 2,589
Reasoning tokens per move
n/a
Aborted on our side (not counted)
0

By phase

PhaseMovesAccuracyCentipawn lossBlunders /100Engine match
Opening1073.119120.040%
Middlegame2n/a170.050%
Endgame0n/an/an/an/a

Chess960

Rating
71395%: 82 to 1,343
Record
0 won, 0 drawn, 2 lost of 2
Accuracy (open positions)
65.7
Accuracy (all moves)
65.7
Average centipawn loss
115
Blunders / mistakes / inaccuracies per 100 moves
23.1 / 23.1 / 38.5
Matched Stockfish's first choice
0%
In Stockfish's top three
23%
Legal on the first try
53.3%
Illegal / invalid
13 of 26 answers
Forfeits
2
Adjudicated games
0
API-equivalent cost per game
n/a
API-equivalent cost per move
n/a
Seconds per move
1,440.0
Model minutes per game
180.0
Tokens per move (in / out)
1,088 / 5,847
Reasoning tokens per move
n/a
Aborted on our side (not counted)
0

By phase

PhaseMovesAccuracyCentipawn lossBlunders /100Engine match
Opening1365.711523.10%
Middlegame0n/an/an/an/a
Endgame0n/an/an/an/a

Games