GLM-5.3 vs GPT-6 Sol (max)

Head-to-head on the Towards AI writing benchmark: same 10 real YouTube scripts, five runs each, scored blind by a three-family judge panel. GLM-5.3 leads overall, 88.7 to 88.0.

GLM-5.3

#15 Elo 2256 · 88.7/100

GPT-6 Sol (max)

#31 Elo 2166 · 88.0/100

Cost / script

$0.114 vs $0.112

Human baseline

90.9 GLM-5.3 falls below it · GPT-6 Sol (max) falls below it

The verdict

This one is closer than the ranks suggest. GLM-5.3 sits at #15 with 2256.0 Elo and 88.73 overall; GPT-6 Sol (max) is at #31 with 2165.6 Elo and 87.95 overall. Their 95% Elo intervals overlap (2220.4 to 2296.8 against 2127.0 to 2200.7), so do not read much into the exact order. Where GLM-5.3 pulls ahead: Hook Strength (89.10 vs 86.91) and Tone & Voice Match (89.05 vs 85.84). GPT-6 Sol (max) still wins on Visual Cue Quality (89.15 vs 83.51) and Slop Score (EQ-Bench + ours) (94.08 vs 92.25), so it is not a clean sweep. The trade is price: GPT-6 Sol (max) costs about $0.112 per article against $0.114 for GLM-5.3, so you pay for the higher rank. GLM-5.3 publishes open weights; GPT-6 Sol (max) does not. Default to GLM-5.3. Reach for GPT-6 Sol (max) when visual cues, slop score, and the lower price matter more to you than the headline rank.

Pick GLM-5.3 for the stronger board result, the hook, and tone.
Pick GPT-6 Sol (max) for visual cues, slop score, and the lower price.

Reading the Elo intervals. Each range estimates uncertainty around one model's Elo. We repeatedly sample from every model's recorded run scores on each task and recalculate the full ranking. Comparing these separate ranges does not test the Elo gap between two models. The tasks stay fixed, so the ranges do not measure how the ranking would change on new tasks.

Metric by metric

Blue bars: GLM-5.3. Orange bars: GPT-6 Sol (max). Same 0–100 scale; the bold bar wins that metric.

Tone & Voice Match19% weight

89.0

85.8

Writing Craft & Clarity13% weight

89.0

87.8

Substance, Accuracy & Value15% weight

88.0

88.9

Continuity & Emotion14% weight

87.7

86.4

YouTube Best Practices12% weight

88.8

87.9

Hook Strength10% weight

89.1

86.9

Length Adherence8% weight

90.4

91.2

Slop Score (EQ-Bench + ours)5% weight

92.2

94.1

Visual Cue Quality4% weight

83.5

89.2

Everything else that differs

	GLM-5.3	GPT-6 Sol (max)
Overall / 100	88.7	88.0
Writing Elo	2256	2166
Run-to-run spread (± overall std)	1.180	1.330
Cost per script (USD)	0.114	0.112
Avg latency (s)	312.8	171.0
Open weights	Yes	No

Full scorecards: GLM-5.3 · GPT-6 Sol (max). How scoring works: methodology.

← All comparisons