Home/Models/Qwen4 Max
Q

Qwen4 Max

Closed API Alibaba Qwen 2026-11-16 1M context

Alibaba's new flagship: 42.9 points, 13th overall and second among Chinese models, behind only Kimi K3.1 Pro

Intelligence v5
42.9
Leaderboard rank
#13
Input / output
$1.50/$4.50
Blended 3:1
$2.25
Output speed
— tok/s
Value
19 pts/$
Capability profile Normalised to the field maximum
Specifications
Vendor
Alibaba Qwen
Released
2026-11-16
Weights
Closed, API only
Context
1,000,000 tokens
Max output
Undisclosed
Parameters
Undisclosed
Input price
$1.50 / 1M
Output price
$4.50 / 1M
Cache hit
$0.075 / 1M
Output speed
Not tested
Intelligence
42.9(v5)

Benchmark scores

How it is measured →

Ranked 13 · bars are scaled to the field maximum for that benchmark.

BenchmarksScoreVs. best in fieldGap
AA Intelligence Index v5 Intelligence index 42.966% −22.1 pts
AA Intelligence Index v4.3.2——not indexed
GPQA Diamond——not indexed
Terminal-Bench 4.0 Terminal-agent tasks 48.572% −18.7 pts
LiveBench Design——not indexed
Vals Index——not indexed

Assessment

Released on 16 November 2026 as the successor to Qwen3.8 Max (0902). The AA Intelligence Index v5.0.1 source value is 42.89, logged here to one decimal as 42.9, which puts it 13th on the board and second among Chinese models, behind only Kimi K3.1 Pro at 46.0. Terminal-Bench 4.0 is recorded at 48.5% from AA's own run (Alibaba's launch post reports the same 48.5%, so the two sets of figures agree this time). Pricing comes from QwenCloud's official model page: $1.5 input / $4.5 output with implicit cache at $0.075; the same page lists a 991K maximum input, 983K in thinking mode and 393K maximum output, recorded here as a 1M context. Four items are left blank with the reason noted: GPQA, LiveBench and Vals read null on the AA source, meaning those evaluations were not run; output speed and time to first answer token have no AA data yet, since no endpoint was measured on launch day; and the parameter count is unpublished, as the widely repeated 2.4T total / 71B active figures have no first-party source, so pt/pa stay empty too. The weights are not open; Alibaba says the smaller Qwen4 models will follow as open weights.

At the same price

The four models closest on blended price, side by side.

Modelv5Blendedtok/sScore gap
GLM-5.3 Open weightsZ.ai 32.0 $2.15 71 -10.9
GPT-6.1 TerraOpenAI 51.3 $2.00 71 +8.4
DeepSeek V4 Pro Open weightsDeepSeek 23.0 $1.98 75 -19.9
Gemini 4 HydrogenGoogle 54.0 $3.00 188 +11.1

Compare side by side

Line it up against other models on price, speed and every benchmark.

The comparison tool lines up to 4 models side by side with switchable normalisation — built for purchasing decisions.

Open the comparison tool → Back to leaderboard

Related benchmarks

Data sources & methodology

Full methodology & sources →

Every number on this page comes from published evaluations. We run no tests of our own and accept no vendor-submitted scores. Main sources:

Artificial Analysis Intelligence Index v5v4.3.2 (separate column)LMArena Text Arena Elo

Evaluation data in the wider industry also comes from public datasets, leaderboard platforms and their evaluation harnesses, private in-house sets, production logs and user feedback, and synthetic data. This site's current rankings use only the parts that are public, independently checkable and clearly defined.