GLM-5.3-Flash vs GPT-6 Luna Max: latency, speed and cost

GLM-5.3-Flash and GPT-6 Luna Max are level on the intelligence index (27.8 and 27.9). See token costs, speed and which to pick.

Metric	GLM-5.3-Flash	GPT-6 Luna (Max)
Intelligence Index	27.8	27.9
Input price (per 1M tokens)	$0.15	$0.10
Output price (per 1M tokens)	$0.50	$0.50
Output speed (tokens/s)	56	121
Time to first token	1.6s	2.0 min
Release date	2026-08-26	2026-09-22

Metric

GLM-5.3-Flash

GPT-6 Luna (Max)

Intelligence Index

27.8

27.9

Input price (per 1M tokens)

$0.15

$0.10

Output price (per 1M tokens)

$0.50

Output speed (tokens/s)

121

Time to first token

1.6s

2.0 min

Release date

2026-08-26

2026-09-22

If you only looked at the intelligence index, GLM-5.3-Flash and GPT-6 Luna (Max) would be hard to tell apart. This page is written automatically from catalogue numbers only: benchmark scores, list prices and measured speed. GLM-5.3-Flash starts answering far sooner, with a time to first token of 1.6 seconds compared with 2.0 minutes for GPT-6 Luna Max.

GPT-6 Luna Max is the faster writer

Measured output speed is 56 tokens per second for GLM-5.3-Flash and 121 tokens per second for GPT-6 Luna Max. GPT-6 Luna Max is ahead by 117% on throughput. Time to first token is 1.6 seconds for GLM-5.3-Flash and 2.0 minutes for GPT-6 Luna Max. The faster writer, GPT-6 Luna Max, gives away too much at the start to win on time: counting the first-token wait, GLM-5.3-Flash finishes first for any reply shorter than about 12,000 tokens. A first-token wait of that length suits background work better than live conversation.

Put together, an 800-token answer would take about 16 seconds from GLM-5.3-Flash and 2.1 minutes from GPT-6 Luna Max, counting the first-token wait.

When they came out

GLM-5.3-Flash came first, on August 26, 2026; GPT-6 Luna Max followed on September 22, 2026. That is close enough to count them as the same generation.

What they cost

Per million tokens, GLM-5.3-Flash charges $0.15 for input and $0.50 for output, while GPT-6 Luna Max charges $0.10 and $0.50. Output rates are close, so the input rate does most of the work in favour of GPT-6 Luna Max.

Take 1 million input tokens and 200,000 output tokens: that comes to $0.25 on GLM-5.3-Flash and $0.20 on GPT-6 Luna Max. For summarising 1,000 documents of 8,000 tokens into 400-token summaries, GPT-6 Luna Max would come in 29% lower: $1.40 for GLM-5.3-Flash against $1.00 for GPT-6 Luna Max. Blended at three input tokens per output token, the effective rates are $0.238 for GLM-5.3-Flash and $0.20 for GPT-6 Luna Max.

Benchmark scores

With 27.8 for GLM-5.3-Flash and 27.9 for GPT-6 Luna Max, the headline index barely separates them. On an aggregate like this, under a point is effectively level.

For context, that places GLM-5.3-Flash 63rd and GPT-6 Luna Max 62nd 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.

Cost per point of intelligence

Dividing the blended price by the intelligence index gives a rough cost per point: $0.0085 for GLM-5.3-Flash and $0.0072 for GPT-6 Luna Max, per million tokens. With scores this close, GPT-6 Luna Max is simply the cheaper way to buy the same capability.

Quick answers

Is GPT-6 Luna Max better than GLM-5.3-Flash?

It depends on what you need: they are level on the intelligence index (27.8 and 27.9); GPT-6 Luna Max is 16% cheaper on blended price; GPT-6 Luna Max streams output 2.2 times as fast; and GLM-5.3-Flash reaches its first token sooner.

Who makes GLM-5.3-Flash and GPT-6 Luna Max?

GLM-5.3-Flash is made by Z AI and GPT-6 Luna Max by OpenAI. GLM-5.3-Flash was released on August 26, 2026 and GPT-6 Luna Max on September 22, 2026.

Which is cheaper, GLM-5.3-Flash or GPT-6 Luna Max?

GPT-6 Luna Max is cheaper on a blended basis. GLM-5.3-Flash lists $0.15 per million input tokens and $0.50 per million output tokens; GPT-6 Luna Max lists $0.10 and $0.50. For 1 million input and 200,000 output tokens that is $0.25 against $0.20.

Which is faster, GLM-5.3-Flash or GPT-6 Luna Max?

GPT-6 Luna Max generates output faster, at 121 tokens per second against 56 tokens per second. GLM-5.3-Flash reaches its first token sooner (1.6 seconds against 2.0 minutes). Counting the first-token wait, GLM-5.3-Flash finishes first for any reply shorter than about 12,000 tokens.

Where do these numbers come from?

The figures are the recorded benchmark, price and speed data in Franklin AI's model catalogue. This page is written from that data automatically; neither model was tested by hand for it.

Final whistle

In one line: GLM-5.3-Flash has more going for it here, but GPT-6 Luna Max is not without an argument. Pick GLM-5.3-Flash if users are waiting on the first word and replies need to finish sooner at typical lengths. Pick GPT-6 Luna Max if cost matters ($0.50 against $0.50 per million output tokens) and you generate very long outputs, beyond about 12,000 tokens.

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-Flash and GPT-6 Luna (Max) pages. The AI Benchmark methodology explains how that data is collected and shown.