Horror: alone in an eerie place at night

Bad Writing Habits

Detects common prose quality anti-patterns in AI-generated creative writing, including passive voice, past progressive overuse, weak dialogue tags, filter words, purple prose, cliches, AI-ism words/adverbs/names, and more.

Creative Writing Hallucination

Performance Score Distribution (Top 20)

Click a model name to view its detail page.

	Score
GPT-5.4 (Reasoning)	88%
GPT-5.4	88%
GPT-5.4 (Reasoning, Low)	86%
GPT-5.4 Mini	86%
Qwen 3.5 397B A17B	86%
GPT-5.4 Mini (Reasoning, Low)	86%
GPT-5.4 Mini (Reasoning)	85%
DeepSeek V3 (2025-03-24)	85%
Gemini 3.1 Pro (Preview)	85%
Claude Opus 4.5	84%
Z.AI GLM 5 Turbo	83%
Claude Sonnet 4.6 (Reasoning)	83%
Claude Opus 4.6 (Reasoning)	83%
GPT-5	83%
GPT-5.1	83%
Mistral Small 4	83%
Z.AI GLM 5	82%
Claude Opus 4.6	82%
Qwen 3.5 Flash	81%
Writer: Palmyra X5	81%

	Score	Cost	Time
DeepSeek V3 (2025-03-24)	85%	$0.0011	25.4s
GPT-5.4 Mini (Reasoning, Low)	86%	$0.013	15.6s
GPT-5.4 Mini (Reasoning)	85%	$0.015	18.0s
GPT-5.4 Mini	86%	$0.013	15.9s
Qwen 3.5 Flash	81%	$0.0015	28.4s
Z.AI GLM 5	82%	$0.0072	59.0s
LFM2 24B	80%	$0.0002	27.7s
Rocinante 12B	81%	$0.0007	21.4s
Mistral Small 4	83%	$0.0012	21.9s
Mistral NeMO	77%	$0.0003	10.0s
Qwen 3.5 397B A17B	86%	$0.017	1.7m
Z.AI GLM 5 Turbo	83%	$0.0066	34.0s
Claude 3.5 Haiku	78%	$0.0022	10.2s
Mistral Small 4 (Reasoning)	80%	$0.0021	28.9s
Qwen 3.5 9B	79%	$0.0006	53.5s
Ministral 3 3B	77%	$0.0003	7.5s
Gemma 3 12B	78%	$0.0002	36.7s
Ministral 8B	78%	$0.0002	11.1s
Qwen3 235B A22B Instruct 2507	81%	$0.0011	1.1m
Mistral Small Creative	80%	$0.0006	9.8s

	Score	Cost	Speed	Stability
GPT-5.4 Mini	86%	$0.013	15.9s	82%
GPT-5.4 Mini (Reasoning)	85%	$0.015	18.0s	83%
GPT-5.4 Mini (Reasoning, Low)	86%	$0.013	15.6s	82%
DeepSeek V3 (2025-03-24)	85%	$0.0011	25.4s	80%
Mistral Small 4	83%	$0.0012	21.9s	78%
Qwen 3.5 Flash	81%	$0.0015	28.4s	79%
Z.AI GLM 5 Turbo	83%	$0.0066	34.0s	78%
GPT-5.4	88%	$0.048	1.4m	86%
Qwen 3.5 397B A17B	86%	$0.017	1.7m	81%
Rocinante 12B	81%	$0.0007	21.4s	74%
Grok 4.1 Fast	80%	$0.0013	27.0s	75%
Grok 4.20 (Beta)	80%	$0.016	15.3s	77%
LFM2 24B	80%	$0.0002	27.7s	75%
Mistral Small Creative	80%	$0.0006	9.8s	73%
Ministral 3 14B	78%	$0.0005	13.1s	74%
Mistral Large 2	81%	$0.0094	25.1s	75%
Z.AI GLM 5	82%	$0.0072	59.0s	76%
Mistral Small 4 (Reasoning)	80%	$0.0021	28.9s	74%
Writer: Palmyra X5	81%	$0.011	22.8s	73%
Ministral 8B	78%	$0.0002	11.1s	73%

Rank	Model	Avg. Cost	Avg. Time	Stability	# 1	# 2	# 3	# 4	# 5	Total
56	GPT-5.4 (Reasoning)	$0.074	2.1m	82%	93	88	87	87	86	88%
8	GPT-5.4	$0.048	1.4m	86%	90	89	89	87	86	88%
30	GPT-5.4 (Reasoning, Low)	$0.055	1.4m	82%	90	89	86	85	82	86%
1	GPT-5.4 Mini	$0.013	15.9s	82%	89	88	85	85	84	86%
9	Qwen 3.5 397B A17B	$0.017	1.7m	81%	89	87	86	86	81	86%
3	GPT-5.4 Mini (Reasoning, Low)	$0.013	15.6s	82%	88	88	86	85	82	86%
2	GPT-5.4 Mini (Reasoning)	$0.015	18.0s	83%	87	86	86	86	82	85%
4	DeepSeek V3 (2025-03-24)	$0.0011	25.4s	80%	90	87	85	82	82	85%
71	Gemini 3.1 Pro (Preview)	$0.076	1.4m	78%	90	86	83	82	82	85%
36	Claude Opus 4.5	$0.056	50.8s	79%	88	87	84	82	80	84%
7	Z.AI GLM 5 Turbo	$0.0066	34.0s	78%	89	83	82	82	80	83%
23	Claude Sonnet 4.6 (Reasoning)	$0.035	50.5s	78%	88	85	84	81	78	83%
64	Claude Opus 4.6 (Reasoning)	$0.075	1.3m	80%	86	83	83	82	82	83%
91	GPT-5	$0.067	2.5m	79%	86	85	82	82	80	83%
57	GPT-5.1	$0.056	1.4m	79%	87	84	83	81	80	83%
5	Mistral Small 4	$0.0012	21.9s	78%	86	84	82	81	80	83%
17	Z.AI GLM 5	$0.0072	59.0s	76%	88	87	84	78	76	82%
66	Claude Opus 4.6	$0.072	1.2m	80%	85	84	83	82	79	82%
6	Qwen 3.5 Flash	$0.0015	28.4s	79%	84	83	83	78	78	81%
19	Writer: Palmyra X5	$0.011	22.8s	73%	90	80	80	80	78	81%
33	Claude Sonnet 4.6	$0.029	41.9s	77%	85	82	81	80	77	81%
21	GPT-4.1	$0.016	32.8s	75%	86	83	81	79	75	81%
10	Rocinante 12B	$0.0007	21.4s	74%	87	83	83	79	72	81%
27	Gemini 2.5 Flash	$0.0055	11.3s	70%	89	84	78	77	76	81%
41	Qwen3 235B A22B Instruct 2507	$0.0011	1.1m	73%	88	83	80	77	75	81%
16	Mistral Large 2	$0.0094	25.1s	75%	87	81	80	79	76	81%
34	Qwen 3 32B	$0.0014	47.4s	72%	87	83	79	76	76	80%
14	Mistral Small Creative	$0.0006	9.8s	73%	86	84	79	78	75	80%
29	GPT-4o, Aug. 6th (temp=1)	$0.017	31.8s	75%	84	82	80	78	76	80%
13	LFM2 24B	$0.0002	27.7s	75%	86	83	83	76	73	80%
12	Grok 4.20 (Beta)	$0.016	15.3s	77%	83	82	80	78	78	80%
11	Grok 4.1 Fast	$0.0013	27.0s	75%	85	80	80	78	78	80%
18	Mistral Small 4 (Reasoning)	$0.0021	28.9s	74%	85	81	80	79	73	80%
26	Mistral Medium 3.1	$0.0043	34.7s	74%	84	83	79	77	76	80%
25	Mistral Large	$0.0089	23.8s	74%	85	80	78	78	77	80%
83	DeepSeek V3.2	$0.0011	3.2m	76%	83	80	79	78	78	80%
46	GPT-5.4 Nano (Reasoning, Low)	$0.0054	20.0s	70%	85	83	77	76	73	79%
59	Claude Sonnet 4.5	$0.027	35.1s	73%	85	80	79	77	73	79%
54	Gemini 2.5 Pro	$0.035	37.1s	75%	82	81	80	77	75	79%
47	Qwen 3.5 9B	$0.0006	53.5s	73%	83	83	82	77	69	79%
113	Claude Opus 4	$0.156	1.2m	75%	82	79	78	78	77	79%
15	Ministral 3 14B	$0.0005	13.1s	74%	82	80	78	77	75	78%
40	o4 Mini High	$0.018	29.8s	75%	81	81	78	76	76	78%
37	Qwen 3.5 35B	$0.0078	27.8s	73%	83	78	78	78	74	78%
45	MiniMax M2.5	$0.0027	33.4s	72%	84	80	78	77	72	78%
31	ByteDance Seed 1.6 Flash	$0.0011	28.6s	73%	83	80	78	76	74	78%
85	GPT-5.2	$0.054	1.5m	76%	80	79	78	77	77	78%
20	Ministral 8B	$0.0002	11.1s	73%	84	79	79	76	72	78%
39	Gemma 3 12B	$0.0002	36.7s	73%	81	81	80	80	69	78%
49	Ministral 3 8B	$0.0004	8.4s	68%	87	80	76	76	71	78%
86	ByteDance Seed 2.0 Lite	$0.013	2.4m	74%	80	80	78	78	73	78%
22	Claude 3.5 Haiku	$0.0022	10.2s	73%	83	81	80	74	72	78%
42	Mistral Large 3	$0.0022	23.7s	71%	84	79	76	76	74	78%
44	Qwen 3.5 122B	$0.011	27.0s	73%	81	80	77	76	74	78%
60	Claude Sonnet 4	$0.027	41.6s	75%	81	80	78	76	75	78%
50	MoonshotAI: Kimi K2.5	$0.010	1.1m	76%	79	79	79	78	74	78%
81	Hermes 3 70B	$0.0006	32.2s	61%	91	85	74	69	69	78%
28	Stealth: Healer Alpha	$0.0000	23.0s	73%	81	79	77	76	75	78%
32	GPT-5.4 Nano	$0.0056	21.3s	74%	81	78	77	77	76	78%
24	Ministral 3 3B	$0.0003	7.5s	73%	81	80	80	77	69	77%
61	Qwen 3.5 27B	$0.0090	46.7s	71%	84	80	80	71	71	77%
63	Gemma 3 27B	$0.0004	47.6s	69%	85	78	75	75	73	77%
43	Mistral NeMO	$0.0003	10.0s	70%	84	83	81	69	68	77%
65	GPT-4o, May 13th (temp=1)	$0.028	15.6s	71%	82	80	76	75	72	77%
58	GPT-5 Mini	$0.0098	41.5s	72%	81	80	77	74	72	77%
38	GPT-5.4 Nano (Reasoning)	$0.0051	18.9s	73%	80	78	78	77	71	77%
72	Claude Haiku 4.5	$0.0096	20.7s	66%	88	80	78	74	65	77%
35	Gemini 2.5 Flash Lite	$0.0007	8.7s	72%	82	78	77	74	72	76%
67	Grok 4.20 (Beta, Reasoning)	$0.026	24.9s	72%	82	78	77	73	72	76%
76	Stealth: Hunter Alpha	$0.0000	1.1m	68%	83	79	75	75	69	76%
70	MiniMax M2.7	$0.0044	56.7s	70%	83	77	76	74	70	76%
79	Hermes 3 405B	$0.0018	1.3m	69%	85	78	77	71	69	76%
55	Grok 4 Fast	$0.0014	22.7s	70%	81	80	78	72	68	76%
52	GPT-4o Mini (temp=1)	$0.0011	28.0s	71%	81	76	76	73	73	76%
82	Aion 2.0	$0.0051	1.3m	69%	81	77	74	74	71	75%
53	Claude 3 Haiku	$0.0021	14.3s	70%	80	78	75	73	71	75%
75	DeepSeek V3.1	$0.0019	1.5m	72%	78	78	77	74	70	75%
48	Ministral 3B	$0.0001	4.2s	71%	78	78	76	76	69	75%
84	DeepSeek V3 (2024-12-26)	$0.0017	45.8s	64%	86	75	73	72	69	75%
97	Cohere Command R+ (Aug. 2024)	$0.021	1.2m	68%	80	79	74	73	69	75%
88	DeepSeek-V2 Chat	$0.0017	39.2s	63%	85	79	71	71	68	75%
104	Z.AI GLM 4.7	$0.0086	2.1m	69%	78	78	73	72	72	75%
107	Grok 4	$0.039	1.4m	70%	77	77	73	73	72	74%
74	o4 Mini	$0.013	21.1s	69%	81	75	74	73	69	74%
69	Qwen 3.5 Plus (2026-02-15)	$0.0052	30.7s	70%	78	74	73	73	73	74%
87	Claude 3.5 Sonnet	$0.041	35.6s	72%	76	75	74	74	71	74%
62	Gemini 3 Flash (Preview)	$0.0064	17.3s	70%	77	76	73	73	71	74%
95	Z.AI GLM 4.6	$0.0055	41.3s	63%	84	75	71	70	70	74%
68	GPT-4.1 Mini	$0.0026	19.2s	69%	79	78	76	71	67	74%
106	Gemini 3 Pro (Preview)	$0.049	50.2s	69%	79	75	74	73	69	74%
51	Arcee AI: Trinity Mini	$0.0003	9.5s	71%	75	75	74	73	71	74%
117	ByteDance Seed 2.0 Mini	$0.0043	4.8m	67%	80	80	78	66	64	74%
98	Claude 3.7 Sonnet	$0.039	47.5s	70%	77	74	73	72	72	74%
96	Llama 3.1 8B	$0.0004	1.4m	67%	79	76	75	69	66	73%
100	Qwen 2.5 72B	$0.0006	31.0s	61%	82	77	69	69	67	73%
73	Arcee AI: Trinity Large (Preview)	$0.0000	26.3s	69%	75	75	75	74	65	73%
105	Llama 3.1 70B	$0.0007	30.9s	59%	84	77	69	68	64	72%
99	Z.AI GLM 4.5	$0.0049	44.0s	64%	79	72	70	70	69	72%
77	Gemma 3 4B	$0.0001	17.2s	67%	77	72	71	70	68	72%
80	Gemini 2.5 Flash Lite (Reasoning)	$0.0022	21.1s	66%	78	72	71	70	68	72%
94	Gemini 2.5 Flash (Reasoning)	$0.013	25.6s	66%	78	76	74	67	64	72%
90	Gemini 3 Flash (Preview, Reasoning)	$0.0097	24.1s	67%	74	73	70	69	68	71%
103	Z.AI GLM 4.7 Flash	$0.0015	1.2m	67%	76	71	71	69	67	71%
78	GPT-4.1 Nano	$0.0006	10.6s	67%	75	72	71	68	67	71%
118	Mistral Small 3.2 24B	$0.0039	4.0m	61%	79	73	72	70	57	70%
92	GPT-4o, Aug. 6th (temp=0)	$0.016	19.1s	68%	72	72	72	67	66	70%
93	Llama 3.1 Nemotron 70B	$0.0022	33.4s	67%	73	72	71	69	65	70%
89	WizardLM 2 8x22b	$0.0016	23.4s	68%	70	69	69	69	66	69%
101	GPT-4o Mini (temp=0)	$0.0011	26.5s	65%	74	71	70	66	63	69%
112	ByteDance Seed 1.6	$0.013	2.5m	66%	71	70	68	67	67	68%
102	Gemini 3.1 Flash Lite (Preview)	$0.0027	8.8s	64%	70	69	66	66	65	67%
111	GPT-5 Nano	$0.0044	1.5m	61%	72	66	66	65	60	66%
108	GPT-4o, May 13th (temp=0)	$0.028	14.0s	61%	71	67	65	63	62	65%
110	Nemotron 3 Super	$0.0000	1.0m	60%	68	68	63	62	61	64%
109	Inception Mercury 2	$0.0029	7.7s	59%	64	62	61	61	58	61%
115	Nemotron 3 Nano	$0.0007	43.9s	57%	65	63	60	59	56	61%
116	Inception Mercury	$0.012	17.6s	54%	67	62	59	55	55	60%
114	Stealth: Aurora Alpha	$0.0000	9.1s	58%	62	59	59	58	58	59%
76.63%

Median	Evaluator	Top 3	Flop 3
74.9%	"Not X but Y" pattern overuse	100Mistral Small 3.2 24B 100Grok 4.20 (Beta) 100GPT-4o, Aug. 6th (temp=0)	0GPT-5 Nano 0GPT-5 Mini 6Nemotron 3 Nano
84.3%	Adverb-first sentence starts	100LFM2 24B 100GPT-5.4 Nano (Reasoning, Low) 100Grok 4.20 (Beta, Reasoning)	0Inception Mercury 17Inception Mercury 2 18Qwen 2.5 72B
80.0%	Adverbs in dialogue tags	100Claude Sonnet 4.6 (Reasoning) 100Mistral Small 3.2 24B 100MiniMax M2.7	20Grok 4.20 (Beta, Reasoning) 40Mistral Large 3 40GPT-4o, Aug. 6th (temp=1)
85.4%	AI-ism adverb frequency	97Qwen 3.5 9B 97ByteDance Seed 1.6 97o4 Mini High	63Claude Sonnet 4.6 68GPT-4.1 Nano 74GPT-4.1 Mini
100.0%	AI-ism character names	100DeepSeek-V2 Chat 100Gemini 3 Flash (Preview, Reasoning) 100MoonshotAI: Kimi K2.5	68Llama 3.1 8B 96Claude Sonnet 4.5
100.0%	AI-ism location names	100GPT-4.1 Mini 100GPT-5 Nano 100GPT-4.1 Nano	96Qwen 3 32B
18.9%	AI-ism word frequency	73GPT-5 69ByteDance Seed 2.0 Mini 64ByteDance Seed 2.0 Lite	0GPT-4.1 Nano 0Gemini 2.5 Flash 0GPT-4o Mini (temp=0)
100.0%	Cliché density	100DeepSeek V3 (2024-12-26) 100GPT-4o, Aug. 6th (temp=1) 100Gemini 3 Flash (Preview)	53Stealth: Aurora Alpha 53Claude 3 Haiku 60GPT-4.1 Nano
40.8%	Dialogue tag variety (said vs. fancy)	100Stealth: Healer Alpha 100Ministral 3 3B 100Claude 3.5 Haiku	0o4 Mini 0Gemini 2.5 Flash (Reasoning) 0Grok 4.20 (Beta)
5.5%	Em-dash & semicolon overuse	100GPT-5.4 Mini 100Mistral NeMO 100Gemini 3.1 Pro (Preview)	0GPT-5 Nano 0Z.AI GLM 4.5 0GPT-4.1 Mini
100.0%	Emotion telling (show vs. tell)	100Arcee AI: Trinity Mini 100Ministral 3B 100Gemma 3 12B	86Inception Mercury 94Llama 3.1 70B 94GPT-4o Mini (temp=0)
89.5%	Filter word density	100Mistral Medium 3.1 100Mistral Small 4 (Reasoning) 100Mistral Small Creative	0Llama 3.1 Nemotron 70B 0Stealth: Aurora Alpha 0Llama 3.1 70B
100.0%	Gibberish response detection	100Claude Sonnet 4.6 (Reasoning) 100GPT-5.4 (Reasoning) 100MiniMax M2.7	36Llama 3.1 8B 79Qwen 2.5 72B 80ByteDance Seed 2.0 Lite
100.0%	Markdown formatting overuse	100GPT-4o, Aug. 6th (temp=1) 100Llama 3.1 Nemotron 70B 100Mistral Large	83Mistral Medium 3.1 86ByteDance Seed 1.6 Flash 86Qwen3 235B A22B Instruct 2507
100.0%	Missing dialogue indicators (quotation marks)	100Stealth: Aurora Alpha 100MiniMax M2.7 100Inception Mercury	77Z.AI GLM 4.6 78Gemma 3 27B 80Qwen 3.5 397B A17B
96.7%	Name drop frequency	100DeepSeek V3.2 100DeepSeek V3.1 100LFM2 24B	49Claude 3 Haiku 50Claude 3.7 Sonnet 54Qwen 2.5 72B
56.1%	Narrator intent-glossing	100o4 Mini High 100DeepSeek V3 (2025-03-24) 100Mistral Medium 3.1	0Stealth: Aurora Alpha 0Nemotron 3 Nano 0Nemotron 3 Super
100.0%	Overuse of "that" (subordinate clause padding)	100Ministral 3 8B 100Claude Sonnet 4.6 (Reasoning) 100GPT-5 Mini	56Inception Mercury 60ByteDance Seed 1.6 65GPT-5 Nano
99.9%	Paragraph length variance	100Gemini 3.1 Pro (Preview) 100Claude Opus 4.6 100Gemini 3 Flash (Preview)	6Arcee AI: Trinity Mini 17Inception Mercury 28Grok 4 Fast
95.8%	Passive voice overuse	100o4 Mini High 100Claude Sonnet 4.5 100Claude 3.5 Haiku	73ByteDance Seed 1.6 75ByteDance Seed 2.0 Mini 81Inception Mercury
79.7%	Past progressive (was/were + -ing) overuse	100Grok 4.1 Fast 100GPT-5.2 100GPT-5	6Gemini 3 Flash (Preview, Reasoning) 10Qwen 3.5 35B 11Qwen 3.5 27B
82.2%	Pronoun-first sentence starts	100GPT-5.4 (Reasoning, Low) 100Z.AI GLM 4.5 100Claude Sonnet 4.5	7Mistral Small 3.2 24B 26Gemini 3.1 Flash Lite (Preview) 28Qwen 3.5 27B
95.0%	Purple prose (modifier overload)	100Nemotron 3 Nano 100Qwen 2.5 72B 100Stealth: Aurora Alpha	79Gemini 2.5 Flash (Reasoning) 79GPT-5.4 (Reasoning) 79GPT-4.1 Nano
100.0%	Repeated phrase echo	100Gemini 3 Flash (Preview, Reasoning) 100GPT-4o, May 13th (temp=1) 100o4 Mini High	—
100.0%	Sentence length variance	100GPT-5.4 Nano 100Z.AI GLM 4.7 100GPT-5.4 Mini (Reasoning)	70GPT-4o, Aug. 6th (temp=0) 71Inception Mercury 81WizardLM 2 8x22b
40.9%	Sentence opener variety	90Claude 3.5 Haiku 84Llama 3.1 8B 79Grok 4	25Mistral Small 3.2 24B 27GPT-5 Nano 30Inception Mercury
58.4%	Subject-first sentence starts	100GPT-5.4 100Writer: Palmyra X5 100Qwen3 235B A22B Instruct 2507	0Stealth: Aurora Alpha 0Inception Mercury 0Inception Mercury 2
49.1%	Subordinate conjunction sentence starts	100GPT-5.4 98Z.AI GLM 4.7 94Writer: Palmyra X5	0Claude 3 Haiku 0Claude Opus 4 0Mistral Small 3.2 24B
56.3%	Technical jargon density	99Qwen 2.5 72B 94DeepSeek V3 (2025-03-24) 94Qwen 3.5 397B A17B	0GPT-4o Mini (temp=0) 0GPT-5 Nano 0Stealth: Aurora Alpha
60.0%	Useless dialogue additions	100Inception Mercury 100Claude Sonnet 4.6 (Reasoning) 100MiniMax M2.7	0GPT-4o, May 13th (temp=0) 0Arcee AI: Trinity Large (Preview) 0GPT-4o Mini (temp=0)

Bad Writing Habits

Horror: alone in an eerie place at night

Performance Score Distribution (Top 20)

Price-Performance Score Distribution (Top 20)

Most Stable Models (Top 20)

Top Overall Models (Top 20)