Skip to content
Leaders Board.
Models/AlibabaOpen weights

Qwen3-235B-A22B-Thinking (Jul 2025)

Ranked on the boards below.
-
Overall rank
-
Podium score
0
Board wins
Jul 25, 2025
Released
·Chat·Reasoning·Agents and tools

Chat

Which model people prefer in blind head-to-head conversations.

27
MultiChallenge
41.2% · of 29
as “qwen3-235b-a22b”
92
Text Arena
1,423 · of 100
as “qwen3-235b-a22b-instruct-2507”

Reasoning

Hard questions with checkable answers: science, math, puzzles, expert exams.

37
MMLU-Pro
84.5% · of 100 · stale since 2026-03-11
as “Qwen3-235B-A22B-Thinking-2507”
63
ARC-AGI-2
1.3% · of 78
as “Qwen3-235b-a22b Instruct (25/07)”
72
GPQA Diamond
80.1% · of 100
as “Qwen3-235B-A22B-Thinking-2507”
85
Epoch ECI
143.9 · of 100
as “Qwen3-235B-A22B-Thinking (Jul 2025)”

Agents and tools

Calling functions and tools, using MCP servers, searching the web.

20
BFCL
52.1% · of 81
as “Qwen3-235B-A22B-Instruct-2507 (Prompt)”

More from Alibaba

Qwen3.8-Flash-NextQwen3.8 2.4T A95BQwen3.8 Max (0902)Qwen Audio 3.0 TTS PlusQwen3.8 27BQwen 3.8 27BQwen 3.8 MaxQwen3.7 Flash

Scores as published by each board on the capture date. Model names and logos belong to their owners; logos via logo.dev.

Weekly: the AI leaderboards with a new #1, Saturday mornings.