Dialogue content preserved

Test: Text Replacement

Avg. Score

93.8%

Scenarios

Overall Performance

Performance Score Distribution (Top 20)

Click a model name to view its detail page.

Rank ▲	Model	Score	Avg. Cost	Avg. Time	Stability
1	Qwen 2.5 72B	99.6%	$0.0003	12.0s	96%
2	Gemini 3 Flash (Preview)	98.6%	$0.0021	3.5s	93%
3	Gemini 2.5 Flash Lite	97.0%	$0.0003	1.9s	91%
4	Qwen 3.5 Plus (2026-02-15)	97.9%	$0.0017	7.8s	92%
5	Writer: Palmyra X5	98.6%	$0.0040	11.7s	93%
6	Mistral Large 3	97.5%	$0.0012	8.5s	91%
7	Claude Haiku 4.5	97.5%	$0.0040	3.8s	91%
8	Gemma 3 4B	96.8%	$0.0001	6.9s	88%
9	Mistral Large	97.5%	$0.0049	8.3s	91%
10	Mistral Large 2	97.5%	$0.0049	8.4s	91%
11	Claude Opus 4.5	100.0%	$0.020	6.0s	100%
12	Claude Opus 4.6	100.0%	$0.020	6.6s	100%
13	Arcee AI: Trinity Large (Preview)	97.9%	$0.0000	26.5s	92%
14	Claude Sonnet 4.6	98.8%	$0.012	5.4s	93%
15	Claude Sonnet 4.5	98.7%	$0.012	5.4s	93%
16	Grok 4 Fast	96.4%	$0.0010	8.5s	86%
17	Claude Sonnet 4	97.9%	$0.012	6.9s	92%
18	Claude 3.7 Sonnet	97.5%	$0.012	6.5s	91%
19	GPT-4o, May 13th (temp=1)	97.1%	$0.012	4.0s	91%
20	GPT-4o, May 13th (temp=0)	97.0%	$0.012	3.8s	91%
21	Gemini 2.5 Flash	95.2%	$0.0017	2.4s	83%
22	GPT-4.1 Mini	94.8%	$0.0012	8.2s	86%
23	GPT-4.1	96.6%	$0.0060	4.8s	82%
24	Mistral Small 3.2 24B	97.3%	$0.0002	5.5s	73%
25	Gemini 3 Flash (Preview, Reasoning)	98.9%	$0.015	24.9s	93%
26	Grok 4.1 Fast	95.7%	$0.0012	15.5s	80%
27	Mistral Medium 3.1	93.8%	$0.0014	6.3s	80%
28	Hermes 3 405B	95.9%	$0.0013	25.3s	82%
29	Gemma 3 12B	93.9%	$0.0001	10.2s	78%
30	Ministral 3 14B	93.2%	$0.0003	4.5s	77%
31	Claude 3.5 Sonnet	97.5%	$0.024	10.6s	91%
32	GPT-4o, Aug. 6th (temp=0)	96.3%	$0.0074	3.0s	73%
33	Llama 3.1 8B	93.0%	$0.0001	11.5s	76%
34	Ministral 3 3B	92.0%	$0.0001	2.3s	74%
35	Mistral Small Creative	91.6%	$0.0002	3.3s	74%
36	DeepSeek V3.2	96.4%	$0.0005	52.3s	82%
37	GPT-5 Mini	96.8%	$0.0074	46.2s	86%
38	Gemma 3 27B	93.0%	$0.0002	18.5s	76%
39	Z.AI GLM 4.6	96.1%	$0.0059	50.0s	86%
40	GPT-5.1	97.7%	$0.021	22.6s	86%
41	Ministral 3B	90.9%	$0.0001	2.4s	72%
42	Llama 3.1 70B	93.0%	$0.0006	27.8s	76%
43	Minimax M2.5	96.1%	$0.0019	1.2m	88%
44	DeepSeek V3 (2024-12-26)	94.6%	$0.0009	17.7s	67%
45	Llama 3.1 Nemotron 70B	91.4%	$0.0016	16.4s	75%
46	Z.AI GLM 4.5	93.0%	$0.0044	36.1s	80%
47	GPT-5.2	93.4%	$0.013	10.9s	78%
48	Gemini 2.5 Flash (Reasoning)	92.9%	$0.0087	15.8s	76%
49	ByteDance Seed 1.6	95.5%	$0.0057	1.0m	84%
50	Grok 4	99.1%	$0.033	41.7s	94%
51	GPT-4o Mini (temp=1)	89.3%	$0.0005	10.5s	70%
52	Ministral 3 8B	88.4%	$0.0002	3.6s	69%
53	Ministral 8B	87.9%	$0.0001	3.7s	71%
54	GPT-4o, Aug. 6th (temp=1)	93.7%	$0.0073	3.1s	63%
55	Gemini 2.5 Pro	98.8%	$0.039	28.9s	93%
56	DeepSeek-V2 Chat	93.4%	$0.0009	19.1s	61%
57	GPT-4o Mini (temp=0)	88.7%	$0.0004	10.7s	69%
58	Arcee AI: Trinity Mini	92.7%	$0.0002	7.9s	55%
59	GPT-4.1 Nano	88.0%	$0.0003	4.1s	66%
60	Z.AI GLM 5	99.3%	$0.014	1.9m	95%
61	Gemini 3 Pro (Preview)	100.0%	$0.053	35.9s	100%
62	ByteDance Seed 1.6 Flash	88.9%	$0.0009	17.2s	61%
63	Aion 2.0	95.2%	$0.0051	1.1m	69%
64	Mistral NeMO	90.2%	$0.0002	2.8s	50%
65	WizardLM 2 8x22b	92.1%	$0.0009	51.4s	65%
66	Claude Opus 4.6 (Reasoning)	100.0%	$0.063	29.0s	100%
67	DeepSeek V3 (2025-03-24)	92.1%	$0.0007	40.7s	56%
68	Gemini 2.5 Flash Lite (Reasoning)	88.4%	$0.0030	26.8s	62%
69	Z.AI GLM 4.7	98.8%	$0.011	2.4m	91%
70	GPT-5 Nano	90.9%	$0.0035	1.3m	75%
71	Claude Opus 4	97.3%	$0.060	9.4s	91%
72	GPT-5	98.0%	$0.039	59.9s	86%
73	o4 Mini High	95.9%	$0.034	52.1s	79%
74	Claude Sonnet 4.6 (Reasoning)	99.6%	$0.064	41.7s	96%
75	o4 Mini	91.6%	$0.019	29.9s	62%
76	Qwen 3.5 397B A17B	98.8%	$0.011	3.0m	91%
77	DeepSeek V3.1	85.0%	$0.0007	37.9s	41%
78	Z.AI GLM 4.7 Flash	87.3%	$0.0023	1.6m	55%
79	Gemini 3.1 Pro (Preview)	100.0%	$0.086	1.3m	100%
80	MoonshotAI: Kimi K2.5	96.2%	$0.018	3.8m	87%
81	Cohere Command R+ (Aug. 2024)	72.0%	$0.0076	33.3s	31%
82	Rocinante 12B	65.0%	$0.0004	9.2s	16%
83	Claude 3 Haiku	64.8%	$0.0010	5.1s	7%
84	Hermes 3 70B	63.9%	$0.0016	2.1m	10%
93.78%

Individual Scenarios

Generic Prompt

▼

Combined: 3rd person past → 1st person present

Model	# 1	# 2	# 3	# 4	# 5	# 6	# 7	Avg ▼
Claude Opus 4.6 (Reasoning)	100	100	100	100	100	100	100	100.0%
Gemini 3.1 Pro (Preview)	100	100	100	100	100	100	100	100.0%
Claude Sonnet 4.6 (Reasoning)	100	100	100	100	100	100	100	100.0%
GPT-5.1	100	100	100	100	100	100	100	100.0%
Claude Opus 4.6	100	100	100	100	100	100	100	100.0%
Qwen 3.5 397B A17B	100	100	100	100	100	100	100	100.0%
Z.AI GLM 5	100	100	100	100	100	100	100	100.0%
Claude Sonnet 4.6	100	100	100	100	100	100	100	100.0%
Gemini 3 Flash (Preview, Reasoning)	100	100	100	100	100	100	100	100.0%
Claude Opus 4.5	100	100	100	100	100	100	100	100.0%
Aion 2.0	100	100	100	100	100	100	100	100.0%
Gemini 3 Pro (Preview)	100	100	100	100	100	100	100	100.0%
Claude Sonnet 4	100	100	100	100	100	100	100	100.0%
GPT-4.1	100	100	100	100	100	100	100	100.0%
Gemini 2.5 Pro	100	100	100	100	100	100	100	100.0%
Claude Sonnet 4.5	100	100	100	100	100	100	100	100.0%
Gemini 2.5 Flash (Reasoning)	100	100	100	100	100	100	100	100.0%
Qwen 3.5 Plus (2026-02-15)	100	100	100	100	100	100	100	100.0%
Mistral Large 3	100	100	100	100	100	100	100	100.0%
GPT-4o, May 13th (temp=0)	100	100	100	100	100	100	100	100.0%
Gemini 3 Flash (Preview)	100	100	100	100	100	100	100	100.0%
Claude Haiku 4.5	100	100	100	100	100	100	100	100.0%
DeepSeek-V2 Chat	100	100	100	100	100	100	100	100.0%
Claude 3.5 Sonnet	100	100	100	100	100	100	100	100.0%
GPT-4o, May 13th (temp=1)	100	100	100	100	100	100	100	100.0%
DeepSeek V3 (2024-12-26)	100	100	100	100	100	100	100	100.0%
Claude 3.7 Sonnet	100	100	100	100	100	100	100	100.0%
GPT-4.1 Mini	100	100	100	100	100	100	100	100.0%
Hermes 3 405B	100	100	100	100	100	100	100	100.0%
GPT-4o, Aug. 6th (temp=1)	100	100	100	100	100	100	100	100.0%
GPT-4o, Aug. 6th (temp=0)	100	100	100	100	100	100	100	100.0%
Mistral Large 2	100	100	100	100	100	100	100	100.0%
DeepSeek V3.1	100	100	100	100	100	100	100	100.0%
DeepSeek V3.2	100	100	100	100	100	100	100	100.0%
DeepSeek V3 (2025-03-24)	100	100	100	100	100	100	100	100.0%
Gemini 2.5 Flash Lite	100	100	100	100	100	100	100	100.0%
Mistral Large	100	100	100	100	100	100	100	100.0%
Writer: Palmyra X5	100	100	100	100	100	100	100	100.0%
Mistral Small 3.2 24B	100	100	100	100	100	100	100	100.0%
Gemma 3 12B	100	100	100	100	100	100	100	100.0%
Gemma 3 27B	100	100	100	100	100	100	100	100.0%
Mistral Medium 3.1	100	100	100	100	100	100	100	100.0%
Qwen 2.5 72B	100	100	100	100	100	100	100	100.0%
Arcee AI: Trinity Large (Preview)	100	100	100	100	100	100	100	100.0%
Mistral Small Creative	100	100	100	100	100	100	100	100.0%
Hermes 3 70B	100	100	100	100	100	100	100	100.0%
GPT-4.1 Nano	100	100	100	100	100	100	100	100.0%
GPT-5 Mini	100	100	100	100	100	100	90	98.6%
MoonshotAI: Kimi K2.5	100	100	100	100	100	100	90	98.6%
o4 Mini High	100	100	100	100	100	100	90	98.6%
o4 Mini	100	100	100	100	100	100	90	98.6%
Grok 4	100	100	100	100	100	100	90	98.6%
Claude Opus 4	100	100	100	100	100	100	90	98.6%
Ministral 3 14B	100	100	100	100	100	100	90	98.6%
Z.AI GLM 4.7	100	100	100	100	100	90	90	97.1%
Ministral 3 3B	100	100	100	100	90	90	90	95.7%
GPT-5	100	100	100	100	100	100	70	95.7%
Grok 4 Fast	100	100	100	100	90	90	90	95.7%
Gemini 2.5 Flash	100	100	100	100	100	100	70	95.7%
Gemma 3 4B	100	100	100	100	90	90	90	95.7%
ByteDance Seed 1.6	100	100	100	100	90	90	80	94.3%
Z.AI GLM 4.7 Flash	100	100	100	90	90	90	90	94.3%
WizardLM 2 8x22b	100	100	100	100	90	90	80	94.3%
Ministral 3B	100	100	100	100	90	90	80	94.3%
Minimax M2.5	100	100	90	90	90	90	90	92.9%
Z.AI GLM 4.6	100	100	90	90	90	90	80	91.4%
GPT-4o Mini (temp=1)	100	90	90	90	90	90	90	91.4%
GPT-4o Mini (temp=0)	90	90	90	90	90	90	90	90.0%
ByteDance Seed 1.6 Flash	100	100	100	100	90	80	40	87.1%
GPT-5.2	100	100	100	90	80	70	70	87.1%
Grok 4.1 Fast	100	100	90	90	90	80	50	85.7%
Gemini 2.5 Flash Lite (Reasoning)	100	100	100	100	70	70	60	85.7%
Mistral NeMO	100	100	80	80	80	80	80	85.7%
Llama 3.1 70B	100	80	80	80	80	80	80	82.9%
Llama 3.1 8B	100	100	90	80	70	70	70	82.9%
Z.AI GLM 4.5	90	90	90	80	70	70	70	80.0%
GPT-5 Nano	90	90	80	70	70	70	70	77.1%
Llama 3.1 Nemotron 70B	90	80	80	80	70	70	70	77.1%
Arcee AI: Trinity Mini	100	100	100	100	100	0	0	71.4%
Ministral 8B	80	80	80	70	70	60	60	71.4%
Ministral 3 8B	80	70	70	70	70	60	50	67.1%
Cohere Command R+ (Aug. 2024)	100	50	20	20	20	10	10	32.9%
Claude 3 Haiku	100	100	0	0	0	0	0	28.6%
Rocinante 12B	90	20	20	20	10	10	0	24.3%

▼

Multi-character gender swap: Priya(F)->Rohan(M), Mara unchanged

Model	# 1	# 2	# 3	# 4	# 5	# 6	# 7	Avg ▼
Claude Opus 4.6 (Reasoning)	100	100	100	100	100	100	100	100.0%
Gemini 3.1 Pro (Preview)	100	100	100	100	100	100	100	100.0%
Claude Sonnet 4.6 (Reasoning)	100	100	100	100	100	100	100	100.0%
GPT-5 Mini	100	100	100	100	100	100	100	100.0%
GPT-5.1	100	100	100	100	100	100	100	100.0%
Claude Opus 4.6	100	100	100	100	100	100	100	100.0%
GPT-5	100	100	100	100	100	100	100	100.0%
Qwen 3.5 397B A17B	100	100	100	100	100	100	100	100.0%
Z.AI GLM 5	100	100	100	100	100	100	100	100.0%
Claude Sonnet 4.6	100	100	100	100	100	100	100	100.0%
MoonshotAI: Kimi K2.5	100	100	100	100	100	100	100	100.0%
ByteDance Seed 1.6	100	100	100	100	100	100	100	100.0%
Gemini 3 Flash (Preview, Reasoning)	100	100	100	100	100	100	100	100.0%
o4 Mini High	100	100	100	100	100	100	100	100.0%
GPT-5.2	100	100	100	100	100	100	100	100.0%
Claude Opus 4.5	100	100	100	100	100	100	100	100.0%
Grok 4.1 Fast	100	100	100	100	100	100	100	100.0%
Aion 2.0	100	100	100	100	100	100	100	100.0%
Z.AI GLM 4.6	100	100	100	100	100	100	100	100.0%
Gemini 3 Pro (Preview)	100	100	100	100	100	100	100	100.0%
Claude Sonnet 4	100	100	100	100	100	100	100	100.0%
Minimax M2.5	100	100	100	100	100	100	100	100.0%
Z.AI GLM 4.7	100	100	100	100	100	100	100	100.0%
GPT-4.1	100	100	100	100	100	100	100	100.0%
Gemini 2.5 Pro	100	100	100	100	100	100	100	100.0%
o4 Mini	100	100	100	100	100	100	100	100.0%
Grok 4	100	100	100	100	100	100	100	100.0%
Claude Sonnet 4.5	100	100	100	100	100	100	100	100.0%
Claude Opus 4	100	100	100	100	100	100	100	100.0%
Gemini 2.5 Flash (Reasoning)	100	100	100	100	100	100	100	100.0%
Z.AI GLM 4.5	100	100	100	100	100	100	100	100.0%
Grok 4 Fast	100	100	100	100	100	100	100	100.0%
Qwen 3.5 Plus (2026-02-15)	100	100	100	100	100	100	100	100.0%
Gemini 2.5 Flash Lite (Reasoning)	100	100	100	100	100	100	100	100.0%
Mistral Large 3	100	100	100	100	100	100	100	100.0%
GPT-4o, May 13th (temp=0)	100	100	100	100	100	100	100	100.0%
Gemini 3 Flash (Preview)	100	100	100	100	100	100	100	100.0%
Claude Haiku 4.5	100	100	100	100	100	100	100	100.0%
DeepSeek-V2 Chat	100	100	100	100	100	100	100	100.0%
Z.AI GLM 4.7 Flash	100	100	100	100	100	100	100	100.0%
Claude 3.5 Sonnet	100	100	100	100	100	100	100	100.0%
GPT-4o, May 13th (temp=1)	100	100	100	100	100	100	100	100.0%
DeepSeek V3 (2024-12-26)	100	100	100	100	100	100	100	100.0%
Claude 3.7 Sonnet	100	100	100	100	100	100	100	100.0%
GPT-4.1 Mini	100	100	100	100	100	100	100	100.0%
Hermes 3 405B	100	100	100	100	100	100	100	100.0%
GPT-4o, Aug. 6th (temp=1)	100	100	100	100	100	100	100	100.0%
GPT-5 Nano	100	100	100	100	100	100	100	100.0%
GPT-4o, Aug. 6th (temp=0)	100	100	100	100	100	100	100	100.0%
Mistral Large 2	100	100	100	100	100	100	100	100.0%
DeepSeek V3.2	100	100	100	100	100	100	100	100.0%
Gemini 2.5 Flash Lite	100	100	100	100	100	100	100	100.0%
Gemini 2.5 Flash	100	100	100	100	100	100	100	100.0%
Mistral Large	100	100	100	100	100	100	100	100.0%
Writer: Palmyra X5	100	100	100	100	100	100	100	100.0%
GPT-4o Mini (temp=1)	100	100	100	100	100	100	100	100.0%
Mistral Small 3.2 24B	100	100	100	100	100	100	100	100.0%
Gemma 3 12B	100	100	100	100	100	100	100	100.0%
Llama 3.1 70B	100	100	100	100	100	100	100	100.0%
GPT-4o Mini (temp=0)	100	100	100	100	100	100	100	100.0%
Gemma 3 27B	100	100	100	100	100	100	100	100.0%
Mistral Medium 3.1	100	100	100	100	100	100	100	100.0%
Qwen 2.5 72B	100	100	100	100	100	100	100	100.0%
Llama 3.1 Nemotron 70B	100	100	100	100	100	100	100	100.0%
Arcee AI: Trinity Large (Preview)	100	100	100	100	100	100	100	100.0%
ByteDance Seed 1.6 Flash	100	100	100	100	100	100	100	100.0%
Mistral Small Creative	100	100	100	100	100	100	100	100.0%
Hermes 3 70B	100	100	100	100	100	100	100	100.0%
Ministral 3 14B	100	100	100	100	100	100	100	100.0%
GPT-4.1 Nano	100	100	100	100	100	100	100	100.0%
Ministral 3 8B	100	100	100	100	100	100	100	100.0%
Claude 3 Haiku	100	100	100	100	100	100	100	100.0%
WizardLM 2 8x22b	100	100	100	100	100	100	100	100.0%
Arcee AI: Trinity Mini	100	100	100	100	100	100	100	100.0%
Cohere Command R+ (Aug. 2024)	100	100	100	100	100	100	100	100.0%
Gemma 3 4B	100	100	100	100	100	100	100	100.0%
Mistral NeMO	100	100	100	100	100	100	100	100.0%
Ministral 8B	100	100	100	100	100	100	100	100.0%
Llama 3.1 8B	100	100	100	100	100	100	100	100.0%
DeepSeek V3 (2025-03-24)	100	100	100	100	100	100	90	98.6%
Ministral 3 3B	100	100	100	100	90	90	90	95.7%
Ministral 3B	100	90	90	90	90	90	90	91.4%
DeepSeek V3.1	100	100	100	100	100	100	0	85.7%
Rocinante 12B	100	100	100	100	100	80	0	82.9%

▼

Passive voice → active voice

Model	# 1	# 2	# 3	# 4	# 5	# 6	# 7	Avg ▼
Claude Opus 4.6 (Reasoning)	100	100	100	100	100	100	100	100.0%
Gemini 3.1 Pro (Preview)	100	100	100	100	100	100	100	100.0%
Claude Sonnet 4.6 (Reasoning)	100	100	100	100	100	100	100	100.0%
Claude Opus 4.6	100	100	100	100	100	100	100	100.0%
GPT-5	100	100	100	100	100	100	100	100.0%
Qwen 3.5 397B A17B	100	100	100	100	100	100	100	100.0%
Claude Sonnet 4.6	100	100	100	100	100	100	100	100.0%
Claude Opus 4.5	100	100	100	100	100	100	100	100.0%
Gemini 3 Pro (Preview)	100	100	100	100	100	100	100	100.0%
Z.AI GLM 4.7	100	100	100	100	100	100	100	100.0%
Mistral Small 3.2 24B	100	100	100	100	100	100	100	100.0%
Qwen 2.5 72B	100	100	100	100	100	100	100	100.0%
GPT-5 Mini	100	100	100	100	100	100	90	98.6%
Z.AI GLM 5	100	100	100	100	100	100	90	98.6%
Grok 4.1 Fast	100	100	100	100	100	100	90	98.6%
GPT-4.1	100	100	100	100	100	100	90	98.6%
Gemini 2.5 Pro	100	100	100	100	100	100	90	98.6%
Gemini 3 Flash (Preview)	100	100	100	100	100	100	90	98.6%
Aion 2.0	100	100	100	100	100	90	90	97.1%
Z.AI GLM 4.7 Flash	100	100	100	100	100	90	90	97.1%
Gemini 3 Flash (Preview, Reasoning)	100	100	100	100	100	90	90	97.1%
DeepSeek-V2 Chat	100	100	100	100	100	90	90	97.1%
DeepSeek V3 (2024-12-26)	100	100	100	100	100	90	90	97.1%
GPT-5 Nano	100	100	100	100	90	90	90	95.7%
GPT-5.1	100	100	100	100	90	90	90	95.7%
Grok 4 Fast	100	100	100	100	90	90	90	95.7%
DeepSeek V3.2	100	100	100	100	90	90	90	95.7%
Grok 4	100	100	100	90	90	90	90	94.3%
Claude Haiku 4.5	100	100	100	90	90	90	90	94.3%
Arcee AI: Trinity Mini	100	100	100	90	90	90	90	94.3%
ByteDance Seed 1.6	100	100	100	90	90	90	90	94.3%
Hermes 3 405B	100	100	100	100	90	90	80	94.3%
DeepSeek V3 (2025-03-24)	100	100	100	100	90	90	80	94.3%
Minimax M2.5	100	100	90	90	90	90	90	92.9%
GPT-4o, May 13th (temp=0)	100	100	90	90	90	90	90	92.9%
MoonshotAI: Kimi K2.5	100	100	90	90	90	90	90	92.9%
Gemma 3 4B	100	100	100	90	90	90	80	92.9%
GPT-4o, May 13th (temp=1)	100	90	90	90	90	90	90	91.4%
ByteDance Seed 1.6 Flash	100	90	90	90	90	90	90	91.4%
GPT-5.2	100	90	90	90	90	90	90	91.4%
Z.AI GLM 4.6	100	90	90	90	90	90	90	91.4%
Z.AI GLM 4.5	100	90	90	90	90	90	90	91.4%
Gemini 2.5 Flash Lite (Reasoning)	100	90	90	90	90	90	90	91.4%
Writer: Palmyra X5	100	90	90	90	90	90	90	91.4%
Arcee AI: Trinity Large (Preview)	100	90	90	90	90	90	90	91.4%
GPT-4.1 Nano	100	100	100	90	90	80	80	91.4%
Claude Sonnet 4	90	90	90	90	90	90	90	90.0%
Claude Sonnet 4.5	90	90	90	90	90	90	90	90.0%
Claude Opus 4	90	90	90	90	90	90	90	90.0%
Qwen 3.5 Plus (2026-02-15)	90	90	90	90	90	90	90	90.0%
Mistral Large 3	90	90	90	90	90	90	90	90.0%
Claude 3.5 Sonnet	90	90	90	90	90	90	90	90.0%
Claude 3.7 Sonnet	90	90	90	90	90	90	90	90.0%
GPT-4.1 Mini	90	90	90	90	90	90	90	90.0%
Mistral Large 2	90	90	90	90	90	90	90	90.0%
Gemini 2.5 Flash Lite	90	90	90	90	90	90	90	90.0%
Gemini 2.5 Flash	90	90	90	90	90	90	90	90.0%
Mistral Large	90	90	90	90	90	90	90	90.0%
GPT-4o Mini (temp=1)	90	90	90	90	90	90	90	90.0%
GPT-4o Mini (temp=0)	90	90	90	90	90	90	90	90.0%
Hermes 3 70B	100	100	90	90	90	80	80	90.0%
o4 Mini High	100	100	100	100	100	90	40	90.0%
Gemma 3 27B	90	90	90	90	90	90	80	88.6%
Gemini 2.5 Flash (Reasoning)	100	90	90	90	90	80	80	88.6%
Gemma 3 12B	90	90	90	90	90	90	80	88.6%
Llama 3.1 8B	100	100	100	90	80	70	70	87.1%
DeepSeek V3.1	100	100	100	100	100	100	10	87.1%
Llama 3.1 70B	90	90	90	90	90	80	80	87.1%
GPT-4o, Aug. 6th (temp=0)	100	100	100	90	90	90	0	81.4%
Llama 3.1 Nemotron 70B	90	90	80	80	80	70	60	78.6%
GPT-4o, Aug. 6th (temp=1)	90	90	90	90	90	90	0	77.1%
o4 Mini	100	90	90	90	90	40	30	75.7%
WizardLM 2 8x22b	100	90	90	90	90	40	0	71.4%
Mistral Medium 3.1	70	70	70	70	70	70	70	70.0%
Ministral 3 14B	70	70	70	70	70	70	70	70.0%
Ministral 3 8B	70	70	70	70	70	70	70	70.0%
Ministral 8B	70	70	70	70	70	70	60	68.6%
Ministral 3 3B	80	70	70	70	60	60	60	67.1%
Ministral 3B	70	70	70	70	60	60	50	64.3%
Mistral Small Creative	70	70	70	60	60	60	60	64.3%
Mistral NeMO	80	80	70	20	0	0	0	35.7%
Rocinante 12B	60	30	20	20	10	0	0	20.0%
Cohere Command R+ (Aug. 2024)	30	20	10	10	10	10	10	14.3%
Claude 3 Haiku	0	0	0	0	0	0	0	0.0%

▼

Tense rewriting: past to present

Model	# 1	# 2	# 3	# 4	# 5	# 6	# 7	Avg ▼
Claude Opus 4.6 (Reasoning)	100	100	100	100	100	100	100	100.0%
Gemini 3.1 Pro (Preview)	100	100	100	100	100	100	100	100.0%
Claude Opus 4.6	100	100	100	100	100	100	100	100.0%
Claude Opus 4.5	100	100	100	100	100	100	100	100.0%
Gemini 3 Pro (Preview)	100	100	100	100	100	100	100	100.0%
Claude Sonnet 4	100	100	100	100	100	100	100	100.0%
Grok 4	100	100	100	100	100	100	100	100.0%
Claude Sonnet 4.5	100	100	100	100	100	100	100	100.0%
Qwen 3.5 Plus (2026-02-15)	100	100	100	100	100	100	100	100.0%
Mistral Large 3	100	100	100	100	100	100	100	100.0%
Claude 3.5 Sonnet	100	100	100	100	100	100	100	100.0%
Claude 3.7 Sonnet	100	100	100	100	100	100	100	100.0%
Mistral Large 2	100	100	100	100	100	100	100	100.0%
Mistral Large	100	100	100	100	100	100	100	100.0%
Writer: Palmyra X5	100	100	100	100	100	100	100	100.0%
Mistral Small 3.2 24B	100	100	100	100	100	100	100	100.0%
Arcee AI: Trinity Large (Preview)	100	100	100	100	100	100	100	100.0%
Ministral 3 14B	100	100	100	100	100	100	100	100.0%
Ministral 3 8B	100	100	100	100	100	100	100	100.0%
Gemma 3 4B	100	100	100	100	100	100	100	100.0%
Mistral NeMO	100	100	100	100	100	100	100	100.0%
Claude Sonnet 4.6 (Reasoning)	100	100	100	100	100	90	90	97.1%
Qwen 2.5 72B	100	100	100	100	100	90	90	97.1%
Ministral 8B	100	100	100	100	100	100	80	97.1%
Z.AI GLM 5	100	100	100	100	90	90	90	95.7%
Gemini 3 Flash (Preview, Reasoning)	100	100	100	100	100	90	80	95.7%
Gemini 2.5 Flash Lite	100	100	100	100	90	90	90	95.7%
Ministral 3 3B	100	100	100	90	90	90	90	94.3%
Z.AI GLM 4.7	100	100	100	100	90	90	70	92.9%
GPT-4o, May 13th (temp=1)	100	100	90	90	90	90	90	92.9%
Llama 3.1 8B	100	100	90	90	90	90	90	92.9%
Gemini 2.5 Pro	100	90	90	90	90	90	90	91.4%
GPT-4.1 Nano	100	100	90	90	90	90	80	91.4%
Qwen 3.5 397B A17B	100	100	90	90	90	90	70	90.0%
Claude Sonnet 4.6	90	90	90	90	90	90	90	90.0%
Claude Opus 4	90	90	90	90	90	90	90	90.0%
GPT-4o, May 13th (temp=0)	90	90	90	90	90	90	90	90.0%
Gemini 3 Flash (Preview)	90	90	90	90	90	90	90	90.0%
Claude Haiku 4.5	90	90	90	90	90	90	90	90.0%
GPT-4o, Aug. 6th (temp=0)	90	90	90	90	90	90	90	90.0%
Llama 3.1 70B	90	90	90	90	90	90	90	90.0%
Mistral Medium 3.1	90	90	90	90	90	90	90	90.0%
Llama 3.1 Nemotron 70B	90	90	90	90	90	90	90	90.0%
Ministral 3B	100	100	90	90	90	80	80	90.0%
Minimax M2.5	100	100	90	90	90	90	70	90.0%
GPT-4o, Aug. 6th (temp=1)	100	90	90	90	90	90	80	90.0%
Mistral Small Creative	100	90	90	90	90	90	70	88.6%
GPT-5	100	100	100	100	80	70	70	88.6%
Z.AI GLM 4.6	100	100	100	80	80	80	70	87.1%
MoonshotAI: Kimi K2.5	100	90	90	90	90	70	70	85.7%
Hermes 3 70B	90	90	90	90	80	80	80	85.7%
GPT-5.1	100	100	100	80	80	70	70	85.7%
Arcee AI: Trinity Mini	100	100	100	100	100	100	0	85.7%
Grok 4.1 Fast	100	100	90	80	70	70	70	82.9%
Grok 4 Fast	100	90	90	80	80	70	70	82.9%
GPT-5 Mini	90	90	90	80	80	70	70	81.4%
DeepSeek-V2 Chat	90	90	90	90	70	70	70	81.4%
GPT-4.1 Mini	90	90	80	80	80	80	70	81.4%
Aion 2.0	100	90	90	80	80	70	50	80.0%
Gemini 2.5 Flash	90	90	80	80	80	80	60	80.0%
Cohere Command R+ (Aug. 2024)	100	100	100	90	80	50	40	80.0%
o4 Mini High	90	90	80	80	80	70	70	80.0%
ByteDance Seed 1.6	90	80	80	80	80	70	70	78.6%
ByteDance Seed 1.6 Flash	80	80	80	80	80	80	70	78.6%
Z.AI GLM 4.5	90	80	80	80	70	70	70	77.1%
Z.AI GLM 4.7 Flash	90	90	80	80	80	60	60	77.1%
o4 Mini	90	90	80	80	70	70	50	75.7%
DeepSeek V3 (2025-03-24)	90	90	90	90	90	70	10	75.7%
DeepSeek V3 (2024-12-26)	90	90	90	90	70	50	50	75.7%
Hermes 3 405B	80	80	80	80	80	70	60	75.7%
DeepSeek V3.2	90	90	80	70	70	70	60	75.7%
GPT-4.1	90	80	70	70	70	70	70	74.3%
WizardLM 2 8x22b	90	90	80	70	70	60	60	74.3%
Gemini 2.5 Flash (Reasoning)	90	90	80	70	70	50	50	71.4%
GPT-4o Mini (temp=1)	80	70	70	70	70	70	70	71.4%
Rocinante 12B	100	100	100	100	80	20	0	71.4%
GPT-5.2	70	70	70	70	70	70	70	70.0%
GPT-4o Mini (temp=0)	70	70	70	70	70	70	70	70.0%
Gemma 3 12B	80	80	80	80	70	50	50	70.0%
GPT-5 Nano	80	70	70	70	70	60	50	67.1%
Gemma 3 27B	90	80	80	60	60	50	50	67.1%
Gemini 2.5 Flash Lite (Reasoning)	90	80	70	60	50	50	50	64.3%
DeepSeek V3.1	70	70	50	50	50	50	0	48.6%
Claude 3 Haiku	0	0	0	0	0	0	0	0.0%

Specific Prompt

▼

Combined: 3rd person past → 1st person present

Model	# 1	# 2	# 3	# 4	# 5	# 6	# 7	Avg ▼
Claude Opus 4.6 (Reasoning)	100	100	100	100	100	100	100	100.0%
Gemini 3.1 Pro (Preview)	100	100	100	100	100	100	100	100.0%
Claude Sonnet 4.6 (Reasoning)	100	100	100	100	100	100	100	100.0%
GPT-5 Mini	100	100	100	100	100	100	100	100.0%
GPT-5.1	100	100	100	100	100	100	100	100.0%
Claude Opus 4.6	100	100	100	100	100	100	100	100.0%
GPT-5	100	100	100	100	100	100	100	100.0%
Qwen 3.5 397B A17B	100	100	100	100	100	100	100	100.0%
Z.AI GLM 5	100	100	100	100	100	100	100	100.0%
Claude Sonnet 4.6	100	100	100	100	100	100	100	100.0%
MoonshotAI: Kimi K2.5	100	100	100	100	100	100	100	100.0%
Gemini 3 Flash (Preview, Reasoning)	100	100	100	100	100	100	100	100.0%
o4 Mini High	100	100	100	100	100	100	100	100.0%
GPT-5.2	100	100	100	100	100	100	100	100.0%
Claude Opus 4.5	100	100	100	100	100	100	100	100.0%
Grok 4.1 Fast	100	100	100	100	100	100	100	100.0%
Aion 2.0	100	100	100	100	100	100	100	100.0%
Z.AI GLM 4.6	100	100	100	100	100	100	100	100.0%
Gemini 3 Pro (Preview)	100	100	100	100	100	100	100	100.0%
Claude Sonnet 4	100	100	100	100	100	100	100	100.0%
Minimax M2.5	100	100	100	100	100	100	100	100.0%
Z.AI GLM 4.7	100	100	100	100	100	100	100	100.0%
GPT-4.1	100	100	100	100	100	100	100	100.0%
Gemini 2.5 Pro	100	100	100	100	100	100	100	100.0%
Grok 4	100	100	100	100	100	100	100	100.0%
Claude Sonnet 4.5	100	100	100	100	100	100	100	100.0%
Claude Opus 4	100	100	100	100	100	100	100	100.0%
Z.AI GLM 4.5	100	100	100	100	100	100	100	100.0%
Qwen 3.5 Plus (2026-02-15)	100	100	100	100	100	100	100	100.0%
Mistral Large 3	100	100	100	100	100	100	100	100.0%
GPT-4o, May 13th (temp=0)	100	100	100	100	100	100	100	100.0%
Gemini 3 Flash (Preview)	100	100	100	100	100	100	100	100.0%
Claude Haiku 4.5	100	100	100	100	100	100	100	100.0%
DeepSeek-V2 Chat	100	100	100	100	100	100	100	100.0%
Claude 3.5 Sonnet	100	100	100	100	100	100	100	100.0%
GPT-4o, May 13th (temp=1)	100	100	100	100	100	100	100	100.0%
Claude 3.7 Sonnet	100	100	100	100	100	100	100	100.0%
Hermes 3 405B	100	100	100	100	100	100	100	100.0%
GPT-4o, Aug. 6th (temp=1)	100	100	100	100	100	100	100	100.0%
GPT-4o, Aug. 6th (temp=0)	100	100	100	100	100	100	100	100.0%
Mistral Large 2	100	100	100	100	100	100	100	100.0%
DeepSeek V3.2	100	100	100	100	100	100	100	100.0%
Gemini 2.5 Flash Lite	100	100	100	100	100	100	100	100.0%
Gemini 2.5 Flash	100	100	100	100	100	100	100	100.0%
Mistral Large	100	100	100	100	100	100	100	100.0%
Writer: Palmyra X5	100	100	100	100	100	100	100	100.0%
Mistral Small 3.2 24B	100	100	100	100	100	100	100	100.0%
Gemma 3 12B	100	100	100	100	100	100	100	100.0%
Mistral Medium 3.1	100	100	100	100	100	100	100	100.0%
Qwen 2.5 72B	100	100	100	100	100	100	100	100.0%
Llama 3.1 Nemotron 70B	100	100	100	100	100	100	100	100.0%
Arcee AI: Trinity Large (Preview)	100	100	100	100	100	100	100	100.0%
Mistral Small Creative	100	100	100	100	100	100	100	100.0%
Ministral 3 14B	100	100	100	100	100	100	100	100.0%
Ministral 3 8B	100	100	100	100	100	100	100	100.0%
Claude 3 Haiku	100	100	100	100	100	100	100	100.0%
WizardLM 2 8x22b	100	100	100	100	100	100	100	100.0%
Arcee AI: Trinity Mini	100	100	100	100	100	100	100	100.0%
Cohere Command R+ (Aug. 2024)	100	100	100	100	100	100	100	100.0%
Gemma 3 4B	100	100	100	100	100	100	100	100.0%
Ministral 3 3B	100	100	100	100	100	100	100	100.0%
Mistral NeMO	100	100	100	100	100	100	100	100.0%
Ministral 3B	100	100	100	100	100	100	100	100.0%
Rocinante 12B	100	100	100	100	100	100	100	100.0%
ByteDance Seed 1.6	100	100	100	100	100	100	90	98.6%
Grok 4 Fast	100	100	100	100	100	100	90	98.6%
Gemma 3 27B	100	100	100	100	100	100	90	98.6%
Llama 3.1 8B	100	100	100	100	100	100	90	98.6%
GPT-4.1 Mini	100	100	100	100	100	90	90	97.1%
GPT-5 Nano	100	100	100	100	100	90	90	97.1%
Ministral 8B	100	100	100	100	100	100	80	97.1%
Llama 3.1 70B	100	100	100	100	100	90	70	94.3%
GPT-4o Mini (temp=1)	100	100	90	90	90	90	90	92.9%
GPT-4o Mini (temp=0)	90	90	90	90	90	90	90	90.0%
Gemini 2.5 Flash Lite (Reasoning)	100	100	100	90	80	80	70	88.6%
DeepSeek V3.1	100	100	100	100	100	100	20	88.6%
o4 Mini	100	100	100	100	100	100	0	85.7%
Gemini 2.5 Flash (Reasoning)	100	100	90	90	80	70	70	85.7%
DeepSeek V3 (2024-12-26)	100	100	100	100	100	100	0	85.7%
DeepSeek V3 (2025-03-24)	100	100	100	100	100	100	0	85.7%
ByteDance Seed 1.6 Flash	100	80	80	80	80	60	10	70.0%
Z.AI GLM 4.7 Flash	100	90	90	80	70	30	10	67.1%
GPT-4.1 Nano	90	80	60	60	60	60	60	67.1%
Hermes 3 70B	90	0	0	0	0	0	0	12.9%

▼

Multi-character gender swap: Priya(F)->Rohan(M), Mara unchanged

Model	# 1	# 2	# 3	# 4	# 5	# 6	# 7	Avg ▼
Claude Opus 4.6 (Reasoning)	100	100	100	100	100	100	100	100.0%
Gemini 3.1 Pro (Preview)	100	100	100	100	100	100	100	100.0%
Claude Sonnet 4.6 (Reasoning)	100	100	100	100	100	100	100	100.0%
GPT-5.1	100	100	100	100	100	100	100	100.0%
Claude Opus 4.6	100	100	100	100	100	100	100	100.0%
GPT-5	100	100	100	100	100	100	100	100.0%
Qwen 3.5 397B A17B	100	100	100	100	100	100	100	100.0%
Z.AI GLM 5	100	100	100	100	100	100	100	100.0%
Claude Sonnet 4.6	100	100	100	100	100	100	100	100.0%
MoonshotAI: Kimi K2.5	100	100	100	100	100	100	100	100.0%
ByteDance Seed 1.6	100	100	100	100	100	100	100	100.0%
o4 Mini High	100	100	100	100	100	100	100	100.0%
GPT-5.2	100	100	100	100	100	100	100	100.0%
Claude Opus 4.5	100	100	100	100	100	100	100	100.0%
Grok 4.1 Fast	100	100	100	100	100	100	100	100.0%
Aion 2.0	100	100	100	100	100	100	100	100.0%
Z.AI GLM 4.6	100	100	100	100	100	100	100	100.0%
Gemini 3 Pro (Preview)	100	100	100	100	100	100	100	100.0%
Claude Sonnet 4	100	100	100	100	100	100	100	100.0%
Minimax M2.5	100	100	100	100	100	100	100	100.0%
Z.AI GLM 4.7	100	100	100	100	100	100	100	100.0%
GPT-4.1	100	100	100	100	100	100	100	100.0%
Gemini 2.5 Pro	100	100	100	100	100	100	100	100.0%
o4 Mini	100	100	100	100	100	100	100	100.0%
Grok 4	100	100	100	100	100	100	100	100.0%
Claude Sonnet 4.5	100	100	100	100	100	100	100	100.0%
Claude Opus 4	100	100	100	100	100	100	100	100.0%
Gemini 2.5 Flash (Reasoning)	100	100	100	100	100	100	100	100.0%
Z.AI GLM 4.5	100	100	100	100	100	100	100	100.0%
Grok 4 Fast	100	100	100	100	100	100	100	100.0%
Qwen 3.5 Plus (2026-02-15)	100	100	100	100	100	100	100	100.0%
Mistral Large 3	100	100	100	100	100	100	100	100.0%
GPT-4o, May 13th (temp=0)	100	100	100	100	100	100	100	100.0%
Gemini 3 Flash (Preview)	100	100	100	100	100	100	100	100.0%
Claude Haiku 4.5	100	100	100	100	100	100	100	100.0%
DeepSeek-V2 Chat	100	100	100	100	100	100	100	100.0%
Z.AI GLM 4.7 Flash	100	100	100	100	100	100	100	100.0%
Claude 3.5 Sonnet	100	100	100	100	100	100	100	100.0%
GPT-4o, May 13th (temp=1)	100	100	100	100	100	100	100	100.0%
DeepSeek V3 (2024-12-26)	100	100	100	100	100	100	100	100.0%
Claude 3.7 Sonnet	100	100	100	100	100	100	100	100.0%
GPT-4.1 Mini	100	100	100	100	100	100	100	100.0%
Hermes 3 405B	100	100	100	100	100	100	100	100.0%
GPT-4o, Aug. 6th (temp=1)	100	100	100	100	100	100	100	100.0%
GPT-5 Nano	100	100	100	100	100	100	100	100.0%
GPT-4o, Aug. 6th (temp=0)	100	100	100	100	100	100	100	100.0%
Mistral Large 2	100	100	100	100	100	100	100	100.0%
DeepSeek V3.2	100	100	100	100	100	100	100	100.0%
DeepSeek V3 (2025-03-24)	100	100	100	100	100	100	100	100.0%
Gemini 2.5 Flash Lite	100	100	100	100	100	100	100	100.0%
Gemini 2.5 Flash	100	100	100	100	100	100	100	100.0%
Mistral Large	100	100	100	100	100	100	100	100.0%
Writer: Palmyra X5	100	100	100	100	100	100	100	100.0%
GPT-4o Mini (temp=1)	100	100	100	100	100	100	100	100.0%
Gemma 3 12B	100	100	100	100	100	100	100	100.0%
Llama 3.1 70B	100	100	100	100	100	100	100	100.0%
GPT-4o Mini (temp=0)	100	100	100	100	100	100	100	100.0%
Gemma 3 27B	100	100	100	100	100	100	100	100.0%
Mistral Medium 3.1	100	100	100	100	100	100	100	100.0%
Qwen 2.5 72B	100	100	100	100	100	100	100	100.0%
Llama 3.1 Nemotron 70B	100	100	100	100	100	100	100	100.0%
Arcee AI: Trinity Large (Preview)	100	100	100	100	100	100	100	100.0%
ByteDance Seed 1.6 Flash	100	100	100	100	100	100	100	100.0%
Mistral Small Creative	100	100	100	100	100	100	100	100.0%
Ministral 3 14B	100	100	100	100	100	100	100	100.0%
GPT-4.1 Nano	100	100	100	100	100	100	100	100.0%
Ministral 3 8B	100	100	100	100	100	100	100	100.0%
Claude 3 Haiku	100	100	100	100	100	100	100	100.0%
WizardLM 2 8x22b	100	100	100	100	100	100	100	100.0%
Arcee AI: Trinity Mini	100	100	100	100	100	100	100	100.0%
Cohere Command R+ (Aug. 2024)	100	100	100	100	100	100	100	100.0%
Gemma 3 4B	100	100	100	100	100	100	100	100.0%
Ministral 3 3B	100	100	100	100	100	100	100	100.0%
Mistral NeMO	100	100	100	100	100	100	100	100.0%
Ministral 8B	100	100	100	100	100	100	100	100.0%
Llama 3.1 8B	100	100	100	100	100	100	100	100.0%
Ministral 3B	100	100	100	100	100	100	100	100.0%
GPT-5 Mini	100	100	100	100	100	100	90	98.6%
Gemini 3 Flash (Preview, Reasoning)	100	100	100	100	100	100	90	98.6%
DeepSeek V3.1	100	100	100	100	100	100	50	92.9%
Gemini 2.5 Flash Lite (Reasoning)	100	100	100	100	100	100	0	85.7%
Mistral Small 3.2 24B	100	100	100	100	100	100	0	85.7%
Rocinante 12B	100	100	100	100	100	70	0	81.4%
Hermes 3 70B	100	100	100	90	0	0	0	55.7%

▼

Passive voice → active voice

Model	# 1	# 2	# 3	# 4	# 5	# 6	# 7	Avg ▼
Claude Opus 4.6 (Reasoning)	100	100	100	100	100	100	100	100.0%
Gemini 3.1 Pro (Preview)	100	100	100	100	100	100	100	100.0%
Claude Sonnet 4.6 (Reasoning)	100	100	100	100	100	100	100	100.0%
GPT-5.1	100	100	100	100	100	100	100	100.0%
Claude Opus 4.6	100	100	100	100	100	100	100	100.0%
GPT-5	100	100	100	100	100	100	100	100.0%
Qwen 3.5 397B A17B	100	100	100	100	100	100	100	100.0%
Z.AI GLM 5	100	100	100	100	100	100	100	100.0%
Claude Sonnet 4.6	100	100	100	100	100	100	100	100.0%
ByteDance Seed 1.6	100	100	100	100	100	100	100	100.0%
Gemini 3 Flash (Preview, Reasoning)	100	100	100	100	100	100	100	100.0%
Claude Opus 4.5	100	100	100	100	100	100	100	100.0%
Grok 4.1 Fast	100	100	100	100	100	100	100	100.0%
Gemini 3 Pro (Preview)	100	100	100	100	100	100	100	100.0%
Z.AI GLM 4.7	100	100	100	100	100	100	100	100.0%
GPT-4.1	100	100	100	100	100	100	100	100.0%
Gemini 2.5 Pro	100	100	100	100	100	100	100	100.0%
o4 Mini	100	100	100	100	100	100	100	100.0%
Grok 4	100	100	100	100	100	100	100	100.0%
Claude Sonnet 4.5	100	100	100	100	100	100	100	100.0%
Claude Opus 4	100	100	100	100	100	100	100	100.0%
Gemini 3 Flash (Preview)	100	100	100	100	100	100	100	100.0%
Hermes 3 405B	100	100	100	100	100	100	100	100.0%
DeepSeek V3.2	100	100	100	100	100	100	100	100.0%
GPT-4o Mini (temp=0)	100	100	100	100	100	100	100	100.0%
Qwen 2.5 72B	100	100	100	100	100	100	100	100.0%
Gemma 3 4B	100	100	100	100	100	100	100	100.0%
Mistral NeMO	100	100	100	100	100	100	100	100.0%
o4 Mini High	100	100	100	100	100	100	90	98.6%
GPT-5.2	100	100	100	100	100	100	90	98.6%
Z.AI GLM 4.6	100	100	100	100	100	100	90	98.6%
Grok 4 Fast	100	100	100	100	100	100	90	98.6%
DeepSeek V3 (2024-12-26)	100	100	100	100	100	100	90	98.6%
GPT-4o, Aug. 6th (temp=0)	100	100	100	100	100	100	90	98.6%
GPT-5 Mini	100	100	100	100	100	90	90	97.1%
Gemini 2.5 Flash Lite (Reasoning)	100	100	100	100	100	90	90	97.1%
Writer: Palmyra X5	100	100	100	100	100	90	90	97.1%
WizardLM 2 8x22b	100	100	100	100	100	90	90	97.1%
Gemini 2.5 Flash (Reasoning)	100	100	100	100	100	90	90	97.1%
ByteDance Seed 1.6 Flash	100	100	100	100	100	90	90	97.1%
MoonshotAI: Kimi K2.5	100	100	100	100	90	90	90	95.7%
Z.AI GLM 4.5	100	100	100	100	90	90	90	95.7%
Claude Haiku 4.5	100	100	100	100	90	90	90	95.7%
Gemini 2.5 Flash	100	100	100	100	90	90	90	95.7%
GPT-4o Mini (temp=1)	100	100	100	100	90	90	90	95.7%
GPT-5 Nano	100	100	100	100	90	90	80	94.3%
Minimax M2.5	100	100	90	90	90	90	90	92.9%
Gemma 3 12B	100	100	90	90	90	90	90	92.9%
Arcee AI: Trinity Mini	100	100	90	90	90	90	90	92.9%
Claude Sonnet 4	100	100	90	90	90	90	90	92.9%
Qwen 3.5 Plus (2026-02-15)	100	100	90	90	90	90	90	92.9%
GPT-4o, May 13th (temp=0)	100	100	90	90	90	90	90	92.9%
GPT-4o, May 13th (temp=1)	100	100	90	90	90	90	90	92.9%
Mistral Small 3.2 24B	100	100	90	90	90	90	90	92.9%
Arcee AI: Trinity Large (Preview)	100	90	90	90	90	90	90	91.4%
Mistral Large 3	90	90	90	90	90	90	90	90.0%
Claude 3.5 Sonnet	90	90	90	90	90	90	90	90.0%
Claude 3.7 Sonnet	90	90	90	90	90	90	90	90.0%
GPT-4.1 Mini	90	90	90	90	90	90	90	90.0%
Mistral Large 2	90	90	90	90	90	90	90	90.0%
Gemini 2.5 Flash Lite	90	90	90	90	90	90	90	90.0%
Mistral Large	90	90	90	90	90	90	90	90.0%
Llama 3.1 70B	90	90	90	90	90	90	90	90.0%
Gemma 3 27B	90	90	90	90	90	90	90	90.0%
Mistral Medium 3.1	90	90	90	90	90	90	90	90.0%
Claude 3 Haiku	90	90	90	90	90	90	90	90.0%
Ministral 3 3B	90	90	90	90	90	90	80	88.6%
Ministral 3B	100	90	90	90	90	80	80	88.6%
DeepSeek V3.1	100	100	100	100	100	100	10	87.1%
Llama 3.1 Nemotron 70B	90	90	90	90	80	80	80	85.7%
Z.AI GLM 4.7 Flash	100	100	100	100	90	90	20	85.7%
Aion 2.0	100	100	100	100	100	90	0	84.3%
Llama 3.1 8B	100	100	90	90	80	80	40	82.9%
GPT-4o, Aug. 6th (temp=1)	100	100	100	100	90	90	0	82.9%
DeepSeek V3 (2025-03-24)	100	100	100	100	100	80	0	82.9%
GPT-4.1 Nano	90	90	80	80	80	70	70	80.0%
Mistral Small Creative	80	80	80	80	80	80	80	80.0%
Ministral 3 14B	80	80	80	80	80	70	70	77.1%
Cohere Command R+ (Aug. 2024)	100	100	70	70	60	60	40	71.4%
Ministral 3 8B	70	70	70	70	70	70	70	70.0%
Ministral 8B	70	70	70	70	70	70	70	70.0%
DeepSeek-V2 Chat	100	100	100	90	90	0	0	68.6%
Rocinante 12B	100	90	90	60	40	0	0	54.3%
Hermes 3 70B	90	80	0	0	0	0	0	24.3%

▼

Tense rewriting: past to present

Model	# 1	# 2	# 3	# 4	# 5	# 6	# 7	Avg ▼
Claude Opus 4.6 (Reasoning)	100	100	100	100	100	100	100	100.0%
Gemini 3.1 Pro (Preview)	100	100	100	100	100	100	100	100.0%
Claude Sonnet 4.6 (Reasoning)	100	100	100	100	100	100	100	100.0%
GPT-5 Mini	100	100	100	100	100	100	100	100.0%
GPT-5.1	100	100	100	100	100	100	100	100.0%
Claude Opus 4.6	100	100	100	100	100	100	100	100.0%
GPT-5	100	100	100	100	100	100	100	100.0%
Qwen 3.5 397B A17B	100	100	100	100	100	100	100	100.0%
Z.AI GLM 5	100	100	100	100	100	100	100	100.0%
Claude Sonnet 4.6	100	100	100	100	100	100	100	100.0%
Gemini 3 Flash (Preview, Reasoning)	100	100	100	100	100	100	100	100.0%
o4 Mini High	100	100	100	100	100	100	100	100.0%
GPT-5.2	100	100	100	100	100	100	100	100.0%
Claude Opus 4.5	100	100	100	100	100	100	100	100.0%
Aion 2.0	100	100	100	100	100	100	100	100.0%
Z.AI GLM 4.6	100	100	100	100	100	100	100	100.0%
Gemini 3 Pro (Preview)	100	100	100	100	100	100	100	100.0%
Claude Sonnet 4	100	100	100	100	100	100	100	100.0%
Minimax M2.5	100	100	100	100	100	100	100	100.0%
Z.AI GLM 4.7	100	100	100	100	100	100	100	100.0%
GPT-4.1	100	100	100	100	100	100	100	100.0%
Gemini 2.5 Pro	100	100	100	100	100	100	100	100.0%
Grok 4	100	100	100	100	100	100	100	100.0%
Claude Sonnet 4.5	100	100	100	100	100	100	100	100.0%
Claude Opus 4	100	100	100	100	100	100	100	100.0%
Gemini 2.5 Flash (Reasoning)	100	100	100	100	100	100	100	100.0%
Z.AI GLM 4.5	100	100	100	100	100	100	100	100.0%
Grok 4 Fast	100	100	100	100	100	100	100	100.0%
Qwen 3.5 Plus (2026-02-15)	100	100	100	100	100	100	100	100.0%
Mistral Large 3	100	100	100	100	100	100	100	100.0%
GPT-4o, May 13th (temp=0)	100	100	100	100	100	100	100	100.0%
Gemini 3 Flash (Preview)	100	100	100	100	100	100	100	100.0%
Claude Haiku 4.5	100	100	100	100	100	100	100	100.0%
DeepSeek-V2 Chat	100	100	100	100	100	100	100	100.0%
Claude 3.5 Sonnet	100	100	100	100	100	100	100	100.0%
GPT-4o, May 13th (temp=1)	100	100	100	100	100	100	100	100.0%
DeepSeek V3 (2024-12-26)	100	100	100	100	100	100	100	100.0%
Claude 3.7 Sonnet	100	100	100	100	100	100	100	100.0%
GPT-4.1 Mini	100	100	100	100	100	100	100	100.0%
GPT-4o, Aug. 6th (temp=1)	100	100	100	100	100	100	100	100.0%
GPT-4o, Aug. 6th (temp=0)	100	100	100	100	100	100	100	100.0%
Mistral Large 2	100	100	100	100	100	100	100	100.0%
DeepSeek V3.2	100	100	100	100	100	100	100	100.0%
DeepSeek V3 (2025-03-24)	100	100	100	100	100	100	100	100.0%
Gemini 2.5 Flash Lite	100	100	100	100	100	100	100	100.0%
Gemini 2.5 Flash	100	100	100	100	100	100	100	100.0%
Mistral Large	100	100	100	100	100	100	100	100.0%
Writer: Palmyra X5	100	100	100	100	100	100	100	100.0%
Mistral Small 3.2 24B	100	100	100	100	100	100	100	100.0%
Gemma 3 12B	100	100	100	100	100	100	100	100.0%
Llama 3.1 70B	100	100	100	100	100	100	100	100.0%
Gemma 3 27B	100	100	100	100	100	100	100	100.0%
Mistral Medium 3.1	100	100	100	100	100	100	100	100.0%
Qwen 2.5 72B	100	100	100	100	100	100	100	100.0%
Llama 3.1 Nemotron 70B	100	100	100	100	100	100	100	100.0%
Arcee AI: Trinity Large (Preview)	100	100	100	100	100	100	100	100.0%
Mistral Small Creative	100	100	100	100	100	100	100	100.0%
Ministral 3 14B	100	100	100	100	100	100	100	100.0%
Ministral 3 8B	100	100	100	100	100	100	100	100.0%
Claude 3 Haiku	100	100	100	100	100	100	100	100.0%
WizardLM 2 8x22b	100	100	100	100	100	100	100	100.0%
Mistral NeMO	100	100	100	100	100	100	100	100.0%
Llama 3.1 8B	100	100	100	100	100	100	100	100.0%
ByteDance Seed 1.6	100	100	100	100	100	100	90	98.6%
Grok 4.1 Fast	100	100	100	100	100	100	90	98.6%
Ministral 8B	100	100	100	100	100	100	90	98.6%
Ministral 3B	100	100	100	100	100	100	90	98.6%
Hermes 3 405B	100	100	100	100	100	90	90	97.1%
MoonshotAI: Kimi K2.5	100	100	100	100	100	90	90	97.1%
o4 Mini	100	100	100	100	100	90	90	97.1%
Arcee AI: Trinity Mini	100	100	100	100	100	90	90	97.1%
GPT-5 Nano	100	100	100	100	90	90	90	95.7%
Ministral 3 3B	100	100	100	90	90	90	90	94.3%
Gemini 2.5 Flash Lite (Reasoning)	100	100	100	100	100	90	70	94.3%
DeepSeek V3.1	100	100	100	100	100	100	30	90.0%
ByteDance Seed 1.6 Flash	100	90	90	90	80	80	80	87.1%
Gemma 3 4B	90	90	90	90	80	80	80	85.7%
Rocinante 12B	100	100	100	100	100	100	0	85.7%
Z.AI GLM 4.7 Flash	100	90	90	80	80	80	20	77.1%
Cohere Command R+ (Aug. 2024)	100	100	90	80	60	60	50	77.1%
GPT-4.1 Nano	80	80	80	70	70	70	70	74.3%
GPT-4o Mini (temp=1)	80	80	70	70	70	70	70	72.9%
GPT-4o Mini (temp=0)	70	70	70	70	70	70	70	70.0%
Hermes 3 70B	100	100	100	0	0	0	0	42.9%

Dialogue content preserved

Overall Performance

Performance Score Distribution (Top 20)

Price-Performance Score Distribution (Top 20)

Most Stable Models (Top 20)

Top Overall Models (Top 20)

Individual Scenarios

Generic Prompt

Combined: 3rd person past → 1st person present

Multi-character gender swap: Priya(F)->Rohan(M), Mara unchanged

Passive voice → active voice

Tense rewriting: past to present

Specific Prompt

Combined: 3rd person past → 1st person present

Multi-character gender swap: Priya(F)->Rohan(M), Mara unchanged

Passive voice → active voice

Tense rewriting: past to present