GLM-5.3 Max vs Qwen3.8 Max 0902: speed, latency and cost

Qwen3.8 Max 0902 leads the intelligence index 33.7 to 31.7. GLM-5.3 Max costs 28% less on blended price. Price per token, response speed and when to pick each.

Metric	GLM-5.3 (Max)	Qwen3.8 Max (0902)
Intelligence Index	31.7	33.7
Input price (per 1M tokens)	$1.40	$2.00
Output price (per 1M tokens)	$4.40	$6.00
Output speed (tokens/s)	74	38
Time to first token	1.7s	3.2s
Release date	2026-08-18	2026-09-02

Metric

GLM-5.3 (Max)

Qwen3.8 Max (0902)

Intelligence Index

31.7

33.7

Input price (per 1M tokens)

$1.40

$2.00

Output price (per 1M tokens)

$4.40

$6.00

Output speed (tokens/s)

Time to first token

1.7s

3.2s

Release date

2026-08-18

2026-09-02

If you only looked at the intelligence index, GLM-5.3 (Max) and Qwen3.8 Max (0902) would be hard to tell apart. The biggest gap is output speed, where GLM-5.3 Max records 74 tokens per second and Qwen3.8 Max 0902 38 tokens per second. No hands-on testing went into this write-up; it summarises the recorded benchmark, price and speed data for both models.

How fast they respond

Once a response is under way, GLM-5.3 Max produces roughly 74 tokens per second against 38 tokens per second for Qwen3.8 Max 0902. On long outputs, that is a real advantage for GLM-5.3 Max. Time to first token is 1.7 seconds for GLM-5.3 Max and 3.2 seconds for Qwen3.8 Max 0902. In an interactive app, GLM-5.3 Max will feel noticeably more responsive.

For a 400-token response, first token plus generation comes to roughly 7.1 seconds on GLM-5.3 Max and 14 seconds on Qwen3.8 Max 0902.

The headline benchmark

On the intelligence index, GLM-5.3 Max scores 31.7 against 33.7 for Qwen3.8 Max 0902. It is a small margin for an index that averages many evaluations. In relative terms, Qwen3.8 Max 0902 scores 6% higher.

For context, that places GLM-5.3 Max 47th and Qwen3.8 Max 0902 44th out of 109 scored models still on offer.

Specialist coding and math indexes are missing for both, so treat the overall score as the only capability signal here.

When they came out

The release dates are August 18, 2026 for GLM-5.3 Max and September 2, 2026 for Qwen3.8 Max 0902, 15 days apart. That is close enough to count them as the same generation.

What they cost

The list prices are $1.40 in and $4.40 out for GLM-5.3 Max, and $2.00 in and $6.00 out for Qwen3.8 Max 0902, per million tokens. Qwen3.8 Max 0902 costs 1.4 times as much per output token.

Take 1 million input tokens and 200,000 output tokens: that comes to $2.28 on GLM-5.3 Max and $3.20 on Qwen3.8 Max 0902. A bigger example, classifying 100,000 short messages at 150 input tokens and 5 output tokens each, costs $23.20 on GLM-5.3 Max and $33.00 on Qwen3.8 Max 0902: a gap of $9.80.

Score per dollar

A blunt value measure, blended price over intelligence score, puts GLM-5.3 Max at $0.068 and Qwen3.8 Max 0902 at $0.089 per point. Qwen3.8 Max 0902 costs 1.3 times as much per point, which is the price of its lead.

Quick answers

Is Qwen3.8 Max 0902 better than GLM-5.3 Max?

Not across the board: Qwen3.8 Max 0902 scores higher on the intelligence index, 33.7 to 31.7; GLM-5.3 Max is 28% cheaper on blended price; and GLM-5.3 Max streams output 1.9 times as fast.

Does GLM-5.3 Max respond faster than Qwen3.8 Max 0902?

GLM-5.3 Max generates output faster, at 74 tokens per second against 38 tokens per second. GLM-5.3 Max reaches its first token sooner (1.7 seconds against 3.2 seconds).

Who makes GLM-5.3 Max and Qwen3.8 Max 0902?

GLM-5.3 Max is made by Z AI and Qwen3.8 Max 0902 by Alibaba. GLM-5.3 Max was released on August 18, 2026 and Qwen3.8 Max 0902 on September 2, 2026.

Is Qwen3.8 Max 0902 cheaper than GLM-5.3 Max?

GLM-5.3 Max is cheaper on a blended basis. GLM-5.3 Max lists $1.40 per million input tokens and $4.40 per million output tokens; Qwen3.8 Max 0902 lists $2.00 and $6.00. For 1 million input and 200,000 output tokens that is $2.28 against $3.20.

Which is newer, GLM-5.3 Max or Qwen3.8 Max 0902?

Qwen3.8 Max 0902 is newer. It was released on September 2, 2026, 15 days after GLM-5.3 Max (August 18, 2026).

Final whistle

The verdict: GLM-5.3 Max is the stronger all-round pick on this data, while Qwen3.8 Max 0902 makes sense in narrower cases. Pick GLM-5.3 Max if long outputs need to finish quickly and cost matters ($4.40 against $6.00 per million output tokens). Pick Qwen3.8 Max 0902 if you need the higher overall score (33.7 against 31.7).

Where the numbers come from

The scores, speed and prices in this comparison come from independent third-party model evaluations that Franklin AI syncs each day, and the article was written from a snapshot of them. For the latest results, open the GLM-5.3 (Max) and Qwen3.8 Max (0902) pages. The AI Benchmark methodology explains how that data is collected and shown.