GLM-5.3 vs Qwen3.8 Flash
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 83.0.
The verdict
GLM-5.3 wins this one, and the board is not close: #15 at 2256.0 Elo and 88.73 overall, against #65, 1698.8 Elo and 82.97 overall for Qwen3.8 Flash. The 95% Elo intervals do not overlap (2220.4 to 2296.8 against 1650.3 to 1754.4). Where GLM-5.3 pulls ahead: Continuity & Emotion (87.74 vs 78.74) and Tone & Voice Match (89.05 vs 81.91). Qwen3.8 Flash still wins on Length Adherence (92.46 vs 90.43), so it is not a clean sweep. The trade is price: Qwen3.8 Flash costs about $0.013 per article against $0.114 for GLM-5.3, so you pay for the higher rank. GLM-5.3 publishes open weights; Qwen3.8 Flash does not. Default to GLM-5.3. Reach for Qwen3.8 Flash when length adherence and the lower price matter more to you than the headline rank.
Pick GLM-5.3 for the stronger board result, continuity, and tone.
Pick Qwen3.8 Flash for length adherence 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: Qwen3.8 Flash. Same 0–100 scale; the bold bar wins that metric.
Everything else that differs
| GLM-5.3 | Qwen3.8 Flash | |
|---|---|---|
| Overall / 100 | 88.7 | 83.0 |
| Writing Elo | 2256 | 1699 |
| Run-to-run spread (± overall std) | 1.180 | 2.420 |
| Cost per script (USD) | 0.114 | 0.013 |
| Avg latency (s) | 312.8 | 341.7 |
| Open weights | Yes | No |
Full scorecards: GLM-5.3 · Qwen3.8 Flash. How scoring works: methodology.
← All comparisons