AlibabaChina· released Sep 2025

Qwen3 Max

Name: Qwen3 Max
Author: Alibaba

Provisional: 7 comparable tests so far. Ranked once it reaches 6 across 3 counting categories.

Alibaba, China. One of 19 models from this lab on the Index.

Third-party publishedLow confidence (thin coverage)

Snapshot September 8, 2026

OpenCharts Index

64.2

Provisional: not enough comparable tests or counting categories to rank yet · 7 comparable tests · 1 counting categories

Drop any single test and the Index lands between 51.6 and 64.2.

Needs more results (1 of 8 so far)

Averaging 20.2 on the 1 it has; the score opens at 4 of 8.

Distance to today's frontier on the hardest tests. Not a measure of intelligence.

What it's known for

Qwen3 Max, in 7 results.

Strongest in math (64.2). Its widest gap is on Vending-Bench 2, 100.0% behind the best published result. No comparable results yet in coding, multimodal and long context. Between 35% and 40% behind the best published results, on average.

Strongest category

Math

Competition and research mathematics, graded on the final answer or the proof.

64.2

MATH Level 597.1% · 99% of best
Mock AIME 2024–202573.3% · 73% of best
FrontierMath (Tiers 1–3)18.9% · 20% of best

Mean of 3 of the category's 5 comparable tests. Counts toward the Index.

Best single result

MATH Level 5

Solve the hardest tier of the MATH competition dataset to an exact final answer.

97.1%

99.0% of the best (98.1%) · #4 of 6.

Multi-step algebra and geometry, graded only on the final answer.

Where it trails

Vending-Bench 2

Run a simulated vending business for a year: ordering, pricing and cash flow. The score is the final balance.

$72

100.0% behind the best published result ($11,182) on a log scale · #44 of 53.

Long-horizon decisions whose consequences compound. It exposes models that drift or forget.

Not measured yet

No comparable result yet in these 3 categories. A blank is a blank, never a zero, and it does not lower the Index.

Coding
Multimodal
Long context

Covered, but not averaged into the Index:

Reasoning1 of 6 tests · needs 3 · mean 67.3
Knowledgesingle-test category, shown beside the Index · 64.5
Agentic1 of 8 tests · needs 4 · mean 0.0
Human preferencesingle-test category, shown beside the Index · 79.4

Every score, with its source

See it at work

What Qwen3 Max was asked to do, and how it did.

Every test it has taken, by category, with the kind of task it faced, who produced the number and where the result landed against the best published one. The examples are original and illustrative, never items from the datasets themselves.

Human preference · mean 79.4 across 1 comparable test · a single-test category, shown beside the Index and never averaged into it

Text ArenaLeaderboard

Answer the same prompt as a rival model; real users vote blind on which answer they prefer.

1,435

40% expected win rate against the board leader (1,507)
#63 of 92

max-preview · board 2026-09-02

Source: Arena

Every result is read against the best published one on its test: percent scores as a share of the best above chance, Arena ratings as the expected win rate against the board leader, open-ended values on a log scale. Sitting a harder exam never lowers a model. Every number links to the publisher that produced it.

Head to head

Qwen3 Max against whoever you pick.

Choose a rival. Every counted test both have taken appears side by side as a share of the best published result on that test, with the category means above. A win is a higher share; a gap under 2 points is a tie, because one item on a 45-problem exam is 2.2 points and publishers' own standard errors are of that order.

Qwen3 Max

provisional·Index 64.2C

0 : 4

wins · 0 ties · 4 shared

GPT-6 Astra

#1·Index 99.0A+

GPT-6 Astra comes out ahead on 4 of the 4 tests both have taken.

Category means

Reasoning
67.3
99.6
Knowledge
64.5
100.0
Coding
—
97.7
Math
64.2
99.8
Agentic
0.0
98.5
Human preference
79.4
—

Shared tests · biggest gaps first

FrontierMath (Tiers 1–3)Math
18.9%
93.7%
−79.8
SimpleQA VerifiedKnowledge
48.7%
75.6%
−35.5
GPQA DiamondReasoning
72.6%
95.8%
−32.7
Mock AIME 2024–2025Math
73.3%
100.0%
−26.7

Bars are distances to the best published result on each test (percent scores above chance, Arena ratings as win rate against the leader, open-ended values on a log scale). Raw values as the publishers report them.

Open GPT-6 Astra

Scores

Every test, every source

The best published run per benchmark, the raw value as the publisher reports it, who produced the number, how close it comes to the best published result on that test, and where the model sits among every model scored on it. Percent scores are read above chance, Arena ratings as a win rate against the board leader, open-ended values on a log scale. A test with too few models to compare against is shown but not counted.

Category means

Reasoning67.3 · 1/6
Knowledgebeside64.5 · 1/1
Coding—
Math64.2 · 3/5
Agentic0.0 · 1/8
Multimodal—
Human preferencebeside79.4 · 1/1
Long context—

Mean · tests taken of the category's comparable tests. A faded category either rests on fewer than half of them or is a single-test category (marked “beside”), which is shown next to the Index and never averaged into it.

GPQA DiamondEpoch-runReasoning

72.6%

67.3% of the best (95.8%), above chance
4th percentile of 78 · n = 198

max-2025-09-23

Source: Epoch AI — AI Benchmarking Hub

SimpleQA VerifiedEpoch-runKnowledge

48.7%

64.5% of the best (75.6%)
63th percentile of 57 · n = 1,000

max-2025-09-23

Source: Epoch AI — AI Benchmarking Hub

Mock AIME 2024–2025Epoch-runMath

73.3%