Claude Fable 5.1 (max) vs GPT-5.6 Sol (ultra)

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. Claude Fable 5.1 (max) leads overall, 89.7 to 88.0.

Claude Fable 5.1 (max)

#1 Elo 2290 · 89.7/100

GPT-5.6 Sol (ultra)

#16 Elo 2145 · 88.0/100

Cost / script

$3.153 vs $0.363

Human baseline

90.2 Claude Fable 5.1 (max) falls below it · GPT-5.6 Sol (ultra) falls below it

The verdict

Claude Fable 5.1 (max effort) has the stronger current result: #1 at 2289.7 Elo and 89.71 overall, compared with GPT-5.6 Sol (ultra) at #16, 2145.0 Elo, and 87.97 overall. Their 95% Elo confidence intervals do not overlap, which supports the directional ordering (2229.00–2346.50 and 2104.90–2175.90). Claude Fable 5.1 (max effort) has its clearest metric edges in Length Adherence (94.83 versus 91.90) and Continuity & Emotion (88.54 versus 85.85). GPT-5.6 Sol (ultra) counters on Visual Cue Quality (88.55 versus 85.53) and Substance, Accuracy & Value (89.01 versus 88.88). At the measured run mix, Claude Fable 5.1 (max effort) costs $3.153 per article versus $0.363 for GPT-5.6 Sol (ultra); GPT-5.6 Sol (ultra) is the cheaper route. Both are closed-weight routes. On the current automated evidence, Claude Fable 5.1 (max effort) is the stronger default; GPT-5.6 Sol (ultra) remains a defensible choice when its specific strengths, price, or deployment profile matter more than the headline rank.

Pick Claude Fable 5.1 (max effort) when you prioritize the stronger current board result, length adherence, and continuity & emotion.
Pick GPT-5.6 Sol (ultra) when you prioritize visual cue quality, substance, accuracy & value, and lower measured cost.

Metric by metric

Blue bars: Claude Fable 5.1 (max). Orange bars: GPT-5.6 Sol (ultra). Same 0–100 scale; the bold bar wins that metric.

Tone & Voice Match19% weight

89.3

86.8

Writing Craft & Clarity13% weight

89.2

87.7

Substance, Accuracy & Value15% weight

88.9

89.0

Continuity & Emotion14% weight

88.5

85.8

YouTube Best Practices12% weight

89.0

86.5

Hook Strength10% weight

90.2

87.7

Length Adherence8% weight

94.8

91.9

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

94.0

93.3

Visual Cue Quality4% weight

85.5

88.5

Everything else that differs

	Claude Fable 5.1 (max)	GPT-5.6 Sol (ultra)
Overall / 100	89.7	88.0
Writing Elo	2290	2145
Run-to-run spread (± overall std)	3.870	1.450
Cost per script (USD)	3.153	0.363
Avg latency (s)	622.5	195.5
Open weights	No	No

Full scorecards: Claude Fable 5.1 (max) · GPT-5.6 Sol (ultra). How scoring works: methodology.

← All comparisons