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)	92%
GPT-5.4 (Reasoning, Low)	90%
GPT-5.4	90%
GPT-5.1	88%
Gemini 3.1 Pro (Preview)	87%
Qwen 3.5 397B A17B	86%
GPT-5	86%
GPT-5.4 Mini (Reasoning)	86%
GPT-5.4 Mini (Reasoning, Low)	86%
Grok 4.20 (Beta)	86%
Claude Sonnet 4.5	86%
Z.AI GLM 5	85%
Grok 4.20 (Beta, Reasoning)	85%
Qwen3 235B A22B Instruct 2507	85%
GPT-5.4 Mini	85%
GPT-5.4 Nano	84%
Writer: Palmyra X5	84%
Z.AI GLM 5 Turbo	84%
Claude Sonnet 4.6 (Reasoning)	84%
Rocinante 12B	84%

	Score	Cost	Time
Rocinante 12B	84%	$0.0006	11.9s
GPT-5.4 Mini	85%	$0.013	15.4s
GPT-5.4 Mini (Reasoning)	86%	$0.014	20.5s
Z.AI GLM 5 Turbo	84%	$0.0060	27.7s
GPT-5.4 Nano	84%	$0.0060	23.9s
GPT-5.4 Mini (Reasoning, Low)	86%	$0.012	15.0s
Mistral Small 4 (Reasoning)	82%	$0.0017	23.6s
Mistral Medium 3.1	83%	$0.0030	25.4s
Writer: Palmyra X5	84%	$0.010	20.4s
Ministral 8B	80%	$0.0002	5.2s
Grok 4.20 (Beta)	86%	$0.017	17.0s
Hermes 3 405B	82%	$0.0020	35.8s
GPT-5.4 Nano (Reasoning, Low)	82%	$0.0060	21.9s
Qwen3 235B A22B Instruct 2507	85%	$0.0008	1.0m
GPT-5.4	90%	$0.040	1.2m
Aion 2.0	83%	$0.0050	1.1m
Qwen 3.5 9B	83%	$0.0018	2.7m
Qwen 3 32B	82%	$0.0012	30.5s
Claude Haiku 4.5	81%	$0.0093	19.7s
Grok 4.1 Fast	81%	$0.0014	31.7s

	Score	Consistency	Stability
GPT-5.4 (Reasoning, Low)	90%	97%	87%
GPT-5.4 (Reasoning)	92%	93%	87%
Gemini 3.1 Pro (Preview)	87%	99%	85%
GPT-5.4	90%	93%	84%
GPT-5.4 Mini (Reasoning)	86%	98%	84%
GPT-5.1	88%	95%	84%
GPT-5.4 Mini	85%	96%	83%
GPT-5.4 Nano	84%	97%	82%
Qwen 3.5 397B A17B	86%	95%	82%
Writer: Palmyra X5	84%	97%	82%
Z.AI GLM 5	85%	96%	82%
GPT-5.4 Mini (Reasoning, Low)	86%	96%	82%
Qwen 3.5 27B	82%	98%	81%
GPT-5.4 Nano (Reasoning)	83%	97%	80%
Qwen 3.5 35B	84%	96%	80%
Cohere Command R+ (Aug. 2024)	81%	98%	80%
MiniMax M2.7	83%	95%	80%
Grok 4.20 (Beta)	86%	93%	80%
Grok 4.20 (Beta, Reasoning)	85%	92%	80%
Qwen3 235B A22B Instruct 2507	85%	95%	80%

	Score	Cost	Speed	Stability
GPT-5.4 Mini (Reasoning)	86%	$0.014	20.5s	84%
GPT-5.4 Mini (Reasoning, Low)	86%	$0.012	15.0s	82%
GPT-5.4 Mini	85%	$0.013	15.4s	83%
GPT-5.4 Nano	84%	$0.0060	23.9s	82%
GPT-5.4 (Reasoning, Low)	90%	$0.041	1.1m	87%
Writer: Palmyra X5	84%	$0.010	20.4s	82%
Z.AI GLM 5	85%	$0.0061	59.0s	82%
Grok 4.20 (Beta)	86%	$0.017	17.0s	80%
Rocinante 12B	84%	$0.0006	11.9s	78%
Qwen3 235B A22B Instruct 2507	85%	$0.0008	1.0m	80%
GPT-5.4 Nano (Reasoning)	83%	$0.0054	24.4s	80%
Mistral Medium 3.1	83%	$0.0030	25.4s	80%
GPT-5.4	90%	$0.040	1.2m	84%
Mistral Small 4 (Reasoning)	82%	$0.0017	23.6s	79%
MiniMax M2.7	83%	$0.0040	55.8s	80%
GPT-5.4 Nano (Reasoning, Low)	82%	$0.0060	21.9s	79%
Qwen 3.5 Flash	83%	$0.0021	44.4s	79%
Qwen 3.5 397B A17B	86%	$0.0084	2.0m	82%
Z.AI GLM 5 Turbo	84%	$0.0060	27.7s	77%
Qwen 3 32B	82%	$0.0012	30.5s	78%

Rank	Model	Avg. Cost	Avg. Time	Stability	# 1	# 2	# 3	# 4	# 5	Total
32	GPT-5.4 (Reasoning)	$0.063	1.9m	87%	96	93	93	91	85	92%
5	GPT-5.4 (Reasoning, Low)	$0.041	1.1m	87%	93	91	90	89	89	90%
13	GPT-5.4	$0.040	1.2m	84%	96	91	90	87	86	90%
41	GPT-5.1	$0.042	2.2m	84%	91	90	88	88	84	88%
70	Gemini 3.1 Pro (Preview)	$0.088	1.6m	85%	88	87	87	87	86	87%
18	Qwen 3.5 397B A17B	$0.0084	2.0m	82%	90	87	87	85	82	86%
94	GPT-5	$0.060	2.9m	79%	91	90	85	84	81	86%
1	GPT-5.4 Mini (Reasoning)	$0.014	20.5s	84%	88	87	86	85	85	86%
2	GPT-5.4 Mini (Reasoning, Low)	$0.012	15.0s	82%	90	86	85	85	84	86%
8	Grok 4.20 (Beta)	$0.017	17.0s	80%	91	87	85	83	82	86%
28	Claude Sonnet 4.5	$0.029	38.6s	79%	90	89	85	82	82	86%
7	Z.AI GLM 5	$0.0061	59.0s	82%	89	85	85	84	84	85%
52	Grok 4.20 (Beta, Reasoning)	$0.051	52.0s	80%	90	87	87	85	78	85%
10	Qwen3 235B A22B Instruct 2507	$0.0008	1.0m	80%	90	87	84	84	82	85%
3	GPT-5.4 Mini	$0.013	15.4s	83%	87	87	86	83	82	85%
4	GPT-5.4 Nano	$0.0060	23.9s	82%	86	86	85	83	82	84%
6	Writer: Palmyra X5	$0.010	20.4s	82%	86	85	85	83	82	84%
19	Z.AI GLM 5 Turbo	$0.0060	27.7s	77%	91	87	85	80	77	84%
47	Claude Sonnet 4.6 (Reasoning)	$0.029	40.8s	76%	90	87	82	82	78	84%
9	Rocinante 12B	$0.0006	11.9s	78%	90	86	85	79	79	84%
37	Qwen 3.5 35B	$0.021	1.2m	80%	87	85	84	81	81	84%
88	Claude Opus 4.6	$0.070	1.2m	78%	88	84	83	83	79	83%
84	Claude Opus 4.5	$0.062	51.5s	77%	89	86	83	80	79	83%
15	MiniMax M2.7	$0.0040	55.8s	80%	86	85	84	83	79	83%
29	Aion 2.0	$0.0050	1.1m	78%	86	85	85	84	76	83%
12	Mistral Medium 3.1	$0.0030	25.4s	80%	85	85	84	84	78	83%
61	Qwen 3.5 9B	$0.0018	2.7m	77%	89	86	85	78	77	83%
38	MiniMax M2.5	$0.0030	1.8m	79%	88	85	84	79	79	83%
11	GPT-5.4 Nano (Reasoning)	$0.0054	24.4s	80%	85	84	83	83	80	83%
17	Qwen 3.5 Flash	$0.0021	44.4s	79%	86	84	83	81	80	83%
108	Claude Opus 4	$0.131	56.4s	78%	86	86	85	79	76	83%
102	Qwen 3.5 27B	$0.052	4.0m	81%	84	83	83	81	81	82%
26	Hermes 3 405B	$0.0020	35.8s	76%	87	86	83	80	75	82%
48	Claude Sonnet 4.6	$0.023	35.5s	75%	87	86	82	81	75	82%
23	LFM2 24B	$0.0001	16.1s	75%	89	82	81	80	79	82%
14	Mistral Small 4 (Reasoning)	$0.0017	23.6s	79%	85	85	84	79	77	82%
72	Gemini 3 Pro (Preview)	$0.047	50.8s	77%	86	85	84	80	75	82%
80	ByteDance Seed 2.0 Lite	$0.013	2.4m	75%	86	85	80	80	78	82%
16	GPT-5.4 Nano (Reasoning, Low)	$0.0060	21.9s	79%	85	84	83	80	78	82%
46	Gemini 2.5 Pro	$0.031	32.1s	78%	86	83	83	78	78	82%
39	o4 Mini High	$0.013	23.0s	75%	88	83	81	81	76	82%
101	Claude Opus 4.6 (Reasoning)	$0.076	1.3m	76%	87	85	82	79	77	82%
27	Mistral Large 3	$0.0023	25.5s	76%	87	83	80	80	80	82%
20	Qwen 3 32B	$0.0012	30.5s	78%	84	84	82	79	79	82%
73	WizardLM 2 8x22b	$0.0018	2.7m	75%	88	84	84	77	74	81%
35	Grok 4.1 Fast	$0.0014	31.7s	75%	86	83	82	82	73	81%
30	ByteDance Seed 1.6 Flash	$0.0009	21.4s	75%	87	81	80	79	79	81%
34	Cohere Command R+ (Aug. 2024)	$0.014	50.8s	80%	82	81	81	80	80	81%
25	Claude Haiku 4.5	$0.0093	19.7s	78%	84	84	84	78	77	81%
24	Gemma 3 27B	$0.0004	36.1s	77%	84	82	80	80	79	81%
33	DeepSeek V3 (2025-03-24)	$0.0009	46.1s	77%	84	83	83	80	75	81%
105	Qwen 3.5 122B	$0.067	2.6m	75%	86	83	82	79	73	81%
56	Llama 3.1 8B	$0.0003	1.2m	73%	87	84	83	80	69	81%
50	Z.AI GLM 4.6	$0.0064	1.2m	76%	84	82	80	79	77	80%
21	Ministral 8B	$0.0002	5.2s	76%	84	83	82	79	74	80%
36	Mistral Small Creative	$0.0005	6.9s	74%	84	84	79	78	77	80%
31	Mistral Small 4	$0.0011	15.2s	76%	83	83	79	79	78	80%
87	MoonshotAI: Kimi K2.5	$0.017	2.9m	77%	83	81	80	79	79	80%
81	GPT-5.2	$0.042	1.2m	78%	83	82	81	78	78	80%
103	ByteDance Seed 2.0 Mini	$0.0045	5.2m	76%	84	81	81	79	76	80%
44	Stealth: Hunter Alpha	$0.0000	1.0m	75%	83	82	81	80	73	80%
42	Mistral Large	$0.0079	21.2s	75%	84	81	80	78	75	80%
40	Stealth: Healer Alpha	$0.0000	21.0s	74%	86	81	81	75	74	80%
22	Claude 3.5 Haiku	$0.0022	8.8s	78%	80	80	80	80	76	79%
53	GPT-4o, Aug. 6th (temp=1)	$0.016	21.1s	75%	83	80	80	76	76	79%
75	Claude Sonnet 4	$0.024	37.5s	73%	83	81	77	77	75	79%
98	Hermes 3 70B	$0.0007	2.7m	69%	89	85	85	71	65	79%
67	GPT-4.1	$0.016	47.2s	73%	83	82	79	76	73	79%
43	Gemini 2.5 Flash Lite	$0.0007	7.9s	72%	85	80	78	76	75	79%
71	DeepSeek V3 (2024-12-26)	$0.0015	49.9s	69%	88	79	79	77	70	78%
74	Qwen 3.5 Plus (2026-02-15)	$0.0053	29.2s	69%	85	83	76	75	74	78%
59	Ministral 3 14B	$0.0004	8.7s	69%	87	81	79	77	67	78%
76	DeepSeek V3.2	$0.0010	1.8m	74%	83	79	78	77	75	78%
45	Ministral 3 8B	$0.0003	5.2s	72%	83	82	80	77	70	78%
64	GPT-4o, May 13th (temp=1)	$0.029	13.6s	75%	81	80	79	77	74	78%
58	o4 Mini	$0.012	22.0s	74%	83	79	78	76	74	78%
54	Mistral Large 2	$0.0074	20.2s	74%	82	81	80	74	73	78%
89	DeepSeek V3.1	$0.0017	1.4m	67%	86	84	77	76	68	78%
65	Gemini 3 Flash (Preview, Reasoning)	$0.013	34.1s	73%	81	80	76	76	76	78%
51	Gemini 2.5 Flash (Reasoning)	$0.0093	18.2s	75%	81	80	80	75	73	78%
57	Z.AI GLM 4.5	$0.0047	41.0s	74%	81	78	78	76	75	78%
55	Gemma 3 12B	$0.0002	34.5s	74%	80	78	76	76	76	77%
62	Z.AI GLM 4.7 Flash	$0.0015	1.1m	75%	79	78	78	76	74	77%
78	Gemini 2.5 Flash	$0.0038	8.4s	67%	84	82	74	73	72	77%
83	Arcee AI: Trinity Large (Preview)	$0.0000	32.3s	67%	86	81	76	71	71	77%
63	GPT-4.1 Mini	$0.0023	17.7s	70%	81	80	75	75	74	77%
82	Z.AI GLM 4.7	$0.0097	42.3s	71%	83	76	76	76	72	77%
49	Ministral 3B	$0.0001	2.3s	73%	79	79	76	75	74	77%
104	Grok 4	$0.047	1.8m	71%	83	76	76	76	73	77%
60	Grok 4 Fast	$0.0014	20.5s	72%	80	80	78	73	71	77%
77	GPT-5 Mini	$0.0089	54.8s	73%	80	79	77	74	72	76%
92	Arcee AI: Trinity Mini	$0.0003	8.6s	61%	86	84	71	71	69	76%
68	Gemini 3 Flash (Preview)	$0.0077	19.5s	72%	80	76	75	75	73	76%
79	GPT-4o Mini (temp=1)	$0.0010	33.8s	71%	79	75	74	74	72	75%
95	Gemini 3.1 Flash Lite (Preview)	$0.0026	8.1s	62%	84	82	71	71	67	75%
66	GPT-4o Mini (temp=0)	$0.0010	36.1s	73%	77	75	75	73	73	75%
85	GPT-4.1 Nano	$0.0006	13.0s	68%	82	74	74	74	67	74%
86	Claude 3 Haiku	$0.0020	16.0s	68%	77	76	72	71	70	73%
69	Mistral NeMO	$0.0003	7.1s	71%	75	74	74	73	71	73%
100	Claude 3.7 Sonnet	$0.034	42.3s	72%	74	74	73	73	72	73%
96	DeepSeek-V2 Chat	$0.0016	55.5s	67%	80	77	76	69	64	73%
107	Claude 3.5 Sonnet	$0.033	30.4s	64%	80	74	71	69	67	72%
90	Ministral 3 3B	$0.0002	3.1s	66%	77	77	73	68	65	72%
91	Qwen 2.5 72B	$0.0007	42.1s	70%	73	73	73	72	68	72%
93	Gemma 3 4B	$0.0002	19.4s	67%	76	75	73	71	64	72%
112	ByteDance Seed 1.6	$0.016	3.0m	68%	74	73	73	69	65	71%
97	Llama 3.1 Nemotron 70B	$0.0023	28.5s	67%	75	74	73	66	65	71%
99	Llama 3.1 70B	$0.0007	19.1s	66%	76	71	71	67	65	70%
118	Mistral Small 3.2 24B	$0.0084	6.6m	57%	84	72	67	65	61	70%
106	Gemini 2.5 Flash Lite (Reasoning)	$0.0024	33.9s	62%	76	74	70	62	61	69%
116	Nemotron 3 Super	$0.0000	4.0m	61%	75	69	67	67	61	68%
114	GPT-4o, Aug. 6th (temp=0)	$0.016	23.0s	56%	78	72	65	63	60	68%
110	GPT-4o, May 13th (temp=0)	$0.035	18.3s	64%	72	69	69	64	63	68%
109	Inception Mercury	$0.0058	9.9s	58%	76	66	64	64	61	66%
111	Stealth: Aurora Alpha	$0.0000	8.7s	59%	67	64	63	62	59	63%
113	Inception Mercury 2	$0.0022	5.7s	59%	66	66	62	60	60	63%
115	GPT-5 Nano	$0.0037	1.3m	60%	64	62	62	61	60	62%
117	Nemotron 3 Nano	$0.0012	2.0m	54%	69	69	63	55	53	62%
78.93%

Median	Evaluator	Top 3	Flop 3
83.9%	"Not X but Y" pattern overuse	100Z.AI GLM 4.5 100GPT-4o Mini (temp=1) 100ByteDance Seed 2.0 Lite	0Gemini 2.5 Flash Lite (Reasoning) 0GPT-5 Nano 20Nemotron 3 Nano
80.0%	Adverb-first sentence starts	100Grok 4.20 (Beta) 100GPT-5.4 (Reasoning, Low) 100Mistral Large	0ByteDance Seed 1.6 3Inception Mercury 13Stealth: Aurora Alpha
100.0%	Adverbs in dialogue tags	100Inception Mercury 2 100LFM2 24B 100DeepSeek V3.2	20GPT-5 Nano 40Mistral Large 3 56Hermes 3 70B
90.5%	AI-ism adverb frequency	100Grok 4.1 Fast 99ByteDance Seed 2.0 Lite 99ByteDance Seed 1.6 Flash	70Claude 3 Haiku 74Claude Sonnet 4.6 75Cohere Command R+ (Aug. 2024)
100.0%	AI-ism character names	100o4 Mini 100Claude Opus 4 100MiniMax M2.5	76Llama 3.1 8B 96Stealth: Hunter Alpha 96ByteDance Seed 2.0 Lite
100.0%	AI-ism location names	100Ministral 3 8B 100DeepSeek V3 (2024-12-26) 100Ministral 3B	—
16.2%	AI-ism word frequency	70ByteDance Seed 2.0 Lite 66GPT-5 65ByteDance Seed 2.0 Mini	0Gemini 2.5 Flash (Reasoning) 0Mistral NeMO 0GPT-4o, Aug. 6th (temp=0)
100.0%	Cliché density	100Llama 3.1 8B 100Qwen 3 32B 100Claude Opus 4	33Inception Mercury 2 40GPT-4o, May 13th (temp=0) 53GPT-4o, Aug. 6th (temp=0)
43.3%	Dialogue tag variety (said vs. fancy)	100MiniMax M2.5 100Qwen 3.5 9B 100LFM2 24B	0GPT-4o, May 13th (temp=0) 0Qwen 2.5 72B 0Hermes 3 70B
36.8%	Em-dash & semicolon overuse	100Qwen 2.5 72B 100Qwen 3.5 9B 100Gemini 2.5 Flash (Reasoning)	0GPT-4o, Aug. 6th (temp=1) 0Mistral Small 4 0Claude Sonnet 4.6
100.0%	Emotion telling (show vs. tell)	100Qwen 3.5 9B 100Mistral Medium 3.1 100GPT-5.4 Mini (Reasoning)	90GPT-4o, May 13th (temp=0) 92Hermes 3 70B 93Claude 3 Haiku
96.7%	Filter word density	100Claude Opus 4.6 100Grok 4.20 (Beta, Reasoning) 100Gemini 3.1 Pro (Preview)	0Llama 3.1 70B 0Nemotron 3 Nano 2Inception Mercury 2
100.0%	Gibberish response detection	100GPT-4o, May 13th (temp=0) 100Qwen 3.5 397B A17B 100Grok 4.1 Fast	0Llama 3.1 8B 80Qwen 3.5 9B 92Rocinante 12B
100.0%	Markdown formatting overuse	100GPT-4.1 Nano 100GPT-4o Mini (temp=1) 100Qwen 3.5 Plus (2026-02-15)	63Ministral 3 3B 69ByteDance Seed 1.6 Flash 72Mistral Large 2
100.0%	Missing dialogue indicators (quotation marks)	100Gemini 3.1 Pro (Preview) 100DeepSeek V3 (2024-12-26) 100Llama 3.1 Nemotron 70B	60Qwen 3.5 Flash 77Claude 3 Haiku 80Llama 3.1 70B
92.4%	Name drop frequency	100GPT-5.1 100ByteDance Seed 2.0 Mini 100Gemini 2.5 Flash Lite (Reasoning)	45Claude 3 Haiku 47Mistral Small 4 49Arcee AI: Trinity Large (Preview)
71.4%	Narrator intent-glossing	100Rocinante 12B 100Grok 4.1 Fast 100GPT-5.4 (Reasoning, Low)	0GPT-5 Nano 0Stealth: Aurora Alpha 0Inception Mercury 2
100.0%	Overuse of "that" (subordinate clause padding)	100Gemma 3 27B 100Mistral Small 4 (Reasoning) 100GPT-5.2	78GPT-5 Nano 79Nemotron 3 Nano 79Llama 3.1 70B
100.0%	Paragraph length variance	100GPT-5.4 (Reasoning, Low) 100Claude Sonnet 4.6 100Mistral Small 4	16Inception Mercury 17Mistral Small 3.2 24B 35Gemini 2.5 Flash
98.4%	Passive voice overuse	100GPT-5.4 100Qwen 3.5 397B A17B 100Grok 4.1 Fast	86Llama 3.1 70B 89DeepSeek V3.1 90Mistral Small 3.2 24B
95.9%	Past progressive (was/were + -ing) overuse	100GPT-5.4 Mini (Reasoning, Low) 100Grok 4.1 Fast 100GPT-5	32Z.AI GLM 4.6 33Z.AI GLM 4.7 41DeepSeek V3.2
87.7%	Pronoun-first sentence starts	100Claude 3.5 Sonnet 100Claude Sonnet 4.5 100GPT-5.4 (Reasoning)	1ByteDance Seed 1.6 17Mistral Small 3.2 24B 17ByteDance Seed 2.0 Mini
96.1%	Purple prose (modifier overload)	100Z.AI GLM 5 100Mistral Small 3.2 24B 100o4 Mini	80ByteDance Seed 2.0 Mini 85Gemma 3 27B 85Gemini 2.5 Flash (Reasoning)
100.0%	Repeated phrase echo	100Qwen 3.5 35B 100Mistral Medium 3.1 100Inception Mercury 2	—
100.0%	Sentence length variance	100Ministral 8B 100Z.AI GLM 4.5 100GPT-5.4 (Reasoning)	66GPT-4o, Aug. 6th (temp=0) 74Inception Mercury 77Qwen 3.5 9B
43.1%	Sentence opener variety	92Claude 3.5 Haiku 91Rocinante 12B 90Grok 4.1 Fast	27GPT-5 Nano 28Inception Mercury 29Mistral Small 3.2 24B
43.7%	Subject-first sentence starts	100Qwen3 235B A22B Instruct 2507 99Rocinante 12B 98Writer: Palmyra X5	0Stealth: Aurora Alpha 0Inception Mercury 0Inception Mercury 2
29.7%	Subordinate conjunction sentence starts	84GPT-4o, Aug. 6th (temp=1) 82Mistral NeMO 76Hermes 3 70B	0DeepSeek V3 (2025-03-24) 0Ministral 3B 0Mistral Small 3.2 24B
64.6%	Technical jargon density	100Qwen 3.5 397B A17B 100o4 Mini High 100Qwen 3.5 35B	0GPT-5 Nano 0Nemotron 3 Nano 7Stealth: Aurora Alpha
64.0%	Useless dialogue additions	100Qwen 3.5 9B 100Claude Opus 4.5 100LFM2 24B	0GPT-4o, May 13th (temp=0) 0Stealth: Aurora Alpha 0Gemma 3 12B

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)