Towards AITowards AIToneBench

Model Comparison

Claude Fable 5 vs GLM-5.3

Overlapping ranges: the best configurations, Claude Fable 5 (max effort, Opus 4.8 fallback) and GLM-5.3 (max), are 23 Elo apart, and their 95% ranges overlap. Among 87 models, each ranked by its best configuration, Claude Fable 5 is #7 and GLM-5.3 is #10. Cost per script: $3.20 for Claude Fable 5 (max effort, Opus 4.8 fallback) and $0.240 for GLM-5.3 (max).

Last updated , from board data generated at 17:52 UTC.

Claude Fable 5
2165 Writing Elo of Claude Fable 5 (max effort, Opus 4.8 fallback), its best configuration#7 of 87 models · 82.9/100 Writing Score · $3.20 per script2 configurations on the board
GLM-5.3
2142 Writing Elo of GLM-5.3 (max), its best configuration#10 of 87 models · 82.4/100 Writing Score · $0.240 per script1 configuration on the board

Picks and Verdict

PickClaude Fable 5

Claude Fable 5 (high)

Pick Claude Fable 5 (high) for slightly stronger continuity and YouTube structure at $1.34 a script and under five minutes, since max effort costs about 2.4 times as much with a range that overlaps high's.

PickGLM-5.3

GLM-5.3 (max)

Pick GLM-5.3 (max) for a draft whose range overlaps Fable's best, at $0.240 a script, about a thirteenth of the price.

Claude Fable 5 ranks above GLM-5.3, the #7 model of 87 against the #10 model, each ranked by its best setting, but the step between them is small. Claude Fable 5 (max effort, Opus 4.8 fallback) leads GLM-5.3 (max) by 23 Elo with overlapping 95% ranges. The bill is another story. GLM costs $0.240 a script. Fable at max effort costs $3.20, about 13 times as much, and takes about twice as long. Claude Fable 5 (high) is the version I'd compare: about five and a half times GLM's price, a bit faster at under five minutes, and 72 Elo behind GLM, though their ranges overlap. Going from high to max effort adds 95 Elo for about 2.4 times the cost, and those two ranges overlap a little. At its best, Fable leads seven of the nine metrics, most on YouTube structure and continuity, by over two points each. GLM's hooks are much stronger, more than five points ahead, and its drafts vary less. I'd start with GLM-5.3 and reach for Fable at high only when a script's structure needs the help.

Disclosure: Claude Fable 5 and GLM-5.3 each come from the same lab as one of the judges (Claude Opus 5.5 and GLM 5.3 Flash). Neither is scored by its own lab alone (how the panel works). The judge analysis measures how much each judge favors its own lab.

How to Read the Writing Score, Writing Elo and 95% Ranges

Writing Score and Writing Elo. The Writing Score (0 to 100) is the weighted average of the nine metric scores: seven from the judges, Length Adherence from code and the Slop Score half from each. The Writing Elo is a rating from pairwise comparisons: on each script, each configuration's average across its drafts is compared with every other configuration's, and gaps within the draft-to-draft variation count as draws. Elo rewards winning often, so the two can put configurations in a different order.

Elo ranges. Each 95% range shows how far a configuration's Elo could move if the drafts had come out slightly differently. It comes from resampling every configuration's recorded draft scores on each reference script and ranking again. Comparing two separate ranges does not test the Elo gap between them: ranges far apart are strong evidence for the order, and overlapping ranges do not mean a tie. The reference scripts stay fixed, so the ranges say nothing about new scripts. The methodology page has the full description.

All Thinking Levels of Both Models

Every ranked configuration of Claude Fable 5 and GLM-5.3, scored on the same 10 scripts by the same four judges.

How to Read the Chart
  • Each row is one configuration: the model at one thinking level. The dot is its Writing Elo, a rating from pairwise comparisons of the scores on each script, and the bar is its 95% range: how far that Elo could move if the drafts had come out slightly differently.
  • Each bar describes one configuration. Overlap alone does not establish a tie or decide their order. Comparing these ranges does not directly test the difference between a pair.
  • The table under the chart lists the same configurations in the same order and adds the Writing Score (the 0 to 100 consensus of the judges and code), cost, time and configuration rank (the place among all configurations on the board).
Claude Fable 5GLM-5.3
ConfigurationWriting Elo95% RangeWriting ScoreCost / ScriptTime / ScriptReasoning TokensConfiguration Rank
Claude Fable 5 (max effort, Opus 4.8 fallback)21652102 to 221882.9$3.2012.5 minN/A#17
GLM-5.3 (max)21422111 to 218082.4$0.2406.1 min46,444#22
Claude Fable 5 (high)20702014 to 211681.9$1.344.9 minN/A#28

Cost is one script at the API list price of the route used, uncached and without judging. Time is the average for one accepted script, retries included. N/A under Reasoning Tokens means the board has no separate reasoning count for that route.

Computed Findings Across Thinking Levels

Each finding is computed from the configurations above. The 95% ranges describe individual configurations; overlap alone establishes neither a tie nor an order for a pair.

Two Chosen Configurations Compared

Pick one configuration of each model to compare their metric scores, consistency across drafts and per-script scores. Each menu starts on the model's best configuration.

Scores by Metric

Claude Fable 5 (max effort, Opus 4.8 fallback)GLM-5.3 (max)Solid bar: higher score in that row

Claude Fable 5 (max effort, Opus 4.8 fallback) scores higher on seven of the nine metrics and GLM-5.3 (max) on two. Weighted by each metric's share of the Writing Score, Continuity & Emotion does the most to put Claude Fable 5 (max effort, Opus 4.8 fallback) ahead, with a 2.3-point lead on that metric. The weight beside each metric is its share of the Writing Score.

Tone & Voice Match19% weight
80.6
79.7
Writing Craft & Clarity13% weight
77.5
78.7
Substance, Accuracy & Value15% weight
84.9
83.2
Continuity & Emotion14% weight
84.4
82.1
YouTube Best Practices12% weight
84.3
81.7
Hook Strength10% weight
78.0
83.4
Length Adherence8% weight
94.7
94.3
Slop Score5% weight
84.6
83.4
Visual Cue Quality4% weight
81.2
79.7

Consistency Across Drafts

Claude Fable 5 (max effort, Opus 4.8 fallback)GLM-5.3 (max)
Score spread across drafts (±)1.961.39
Lowest script score81.2 Script 880.5 Script 10
Highest script score86.4 Script 784.1 Script 7
Range across scripts5.23.6

Spread is how much the Writing Score moves between a configuration's drafts, and lower is steadier. Green marks the better number in each row.

Scores by Script

Writing Score on each reference script. Gap is Claude Fable 5 (max effort, Opus 4.8 fallback) minus GLM-5.3 (max), and the higher score is in bold.

ScriptClaude Fable 5 (max effort, Opus 4.8 fallback)GLM-5.3 (max)Gap
Script 1 opinion / warning explainer83.182.1+1.0
Script 2 news-analysis / skeptical explainer83.782.5+1.2
Script 3 personal roadmap / opinion81.681.5+0.1
Script 4 short explainer81.681.0+0.6
Script 5 news-analysis / opinion explainer82.582.50.0
Script 6 founder announcement / personal origin story82.982.90.0
Script 7 personal technical walkthrough / agentic workflow case study86.484.1+2.3
Script 8 product release announcement / personal observation81.282.7-1.5
Script 9 career guide / hiring analysis83.284.0-0.8
Script 10 engineering process walkthrough / presentation adaptation82.880.5+2.3

Compare With Another Model

Pick a second model. A pair with its own written match-up opens that page, and any other pair opens in the comparison tool.

Nearby Match-Ups by Best Writing Elo

Every configuration wrote five drafts for each of the same 10 briefs, from the same prompt. Generation was closed-book, with the prompt as the only source and no web search or other tools. Each judge scores every draft once: Claude Opus 5.5 at medium effort through the Claude Code subscription, GPT-6.1 Sol at medium reasoning through the Codex CLI, GLM 5.3 Flash through OpenRouter, and Jev 1.13, a typed judge from TypeSafe that answers each metric as a multiple-choice question. The channel's finished script is their reference for the voice, and it gets no score on the board itself.

← All comparisons