WizardLM 2 8x22b - NC Bench

microsoft/wizardlm-2-8x22b

WizardLM 2 8x22b

via OpenRouter

Release Date

Apr 15th, 2024

Context Size

65k

Reasoning

Benchmark Cost

$2.37

Speed

206.8 tok/s

Subcategories

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.

Scenario	#1	#2	#3	#4	#5	Total
Detailed Writing Rules
Fantasy: entering an ancient ruin Creative Writing Hallucination	80	79	76	75	75	77%
Horror: alone in an eerie place at night Creative Writing Hallucination	87	84	84	81	81	83%
Literary fiction: old friends reunite Creative Writing Hallucination	85	85	84	82	82	84%
Mystery: examining a crime scene Creative Writing Hallucination	90	87	85	83	82	85%
Romance: separated couple reunites Creative Writing Hallucination	85	84	84	81	81	83%
Thriller: chase through city streets Creative Writing Hallucination	90	83	82	82	79	83%
Detailed Writing Rules						82.57%
genre
Fantasy: entering an ancient ruin Creative Writing Hallucination	72	70	66	62	60	66%
Horror: alone in an eerie place at night Creative Writing Hallucination	70	69	69	69	66	69%
Literary fiction: old friends reunite Creative Writing Hallucination	82	77	76	70	69	75%
Mystery: examining a crime scene Creative Writing Hallucination	78	77	76	69	66	73%
Romance: separated couple reunites Creative Writing Hallucination	69	68	67	66	65	67%
Thriller: chase through city streets Creative Writing Hallucination	90	82	80	76	62	78%
genre						71.32%
Novelcrafter Default Prompt
Fantasy: entering an ancient ruin Creative Writing Hallucination	83	80	77	75	70	77%
Horror: alone in an eerie place at night Creative Writing Hallucination	88	84	84	77	74	81%
Literary fiction: old friends reunite Creative Writing Hallucination	81	80	80	79	79	80%
Mystery: examining a crime scene Creative Writing Hallucination	85	82	81	81	76	81%
Romance: separated couple reunites Creative Writing Hallucination	86	82	80	79	79	81%
Thriller: chase through city streets Creative Writing Hallucination	85	83	81	80	75	81%
Novelcrafter Default Prompt						80.18%
78.02%

Codex Extraction

Evaluates a model's ability to extract structured codex entries (characters, locations, objects, lore) from prose passages and return them as well-formed XML.

Scenario	#1	#2	#3	#4	#5	Total
Long: The Spire of Echoes (Dense) Tooling Reasoning	97	95	95	93	89	94%
Medium: The Hollow (Inferred) Tooling Reasoning	96	96	94	93	93	94%
Medium: Through the Thornveil (Scattered) Tooling Reasoning	96	92	92	89	88	91%
Short: The Rusty Lantern (Explicit) Tooling Reasoning	99	98	96	93	89	95%
93.62%

Codex Red Herring (False Positive Detection)

Tests whether models correctly report "no violations" when a codex is fully consistent with the prose passage. Models that hallucinate false violations (false positives) fail. Uses a 2×2 matrix of text length × codex size, with bare and detailed-entry variants.

Scenario	#1	#2	#3	#4	#5	#6	#7	#8	#9	#10	Total
basic entries
Long text (~1594 words), big codex (51 entries) Hallucination	35	28	27	27	26	25	25	25	13	1	23%
Long text (~1594 words), small codex (11 entries) Hallucination	100	50	50	50	30	8	6	6	5	4	31%
Short text (~524 words), big codex (51 entries) Hallucination	100	50	50	50	42	42	25	25	25	25	43%
Short text (~524 words), small codex (11 entries) Hallucination	50	50	33	25	25	25	17	13	8	6	25%
basic entries											30.64%
detailed entries
Long text (~1594 words), big codex (51 detailed entries) Hallucination	100	100	38	26	17	10	7	4	3	3	31%
Long text (~1594 words), small codex (11 detailed entries) Hallucination	50	31	29	29	29	28	28	13	5	4	25%
Short text (~524 words), big codex (51 detailed entries) Hallucination	100	100	50	42	28	25	25	25	17	1	41%
Short text (~524 words), small codex (11 detailed entries) Hallucination	42	25	25	25	25	13	13	10	8	4	19%
detailed entries											28.84%
29.74%

Codex Violation Detection

Detects factual inconsistencies between a story bible and prose passages. The model must output structured XML identifying each violation with paragraph number and substring.

Scenario	#1	#2	#3	#4	#5	#6	#7	#8	#9	#10	Total
matrix
Large codex (40 entries), long passage (1,019 words) Tooling Reasoning	55	49	48	46	46	43	43	42	39	39	45%
Large codex (40 entries), short passage (165 words) Tooling Reasoning	3	3	3	3	0	0	0	0	0	0	1%
Small codex (7 entries), long passage (734 words) Tooling Reasoning	33	33	33	33	23	21	17	17	3	3	22%
Small codex (7 entries), short passage (165 words) Tooling Reasoning	70	53	22	17	3	3	3	3	0	0	18%
matrix											21.41%
tiers
5 codex entries Tooling Reasoning	3	3	3	3	3	3	0	0	0	0	2%
10 codex entries Tooling Reasoning	100	94	92	86	86	79	79	72	72	0	76%
20 codex entries Tooling Reasoning	82	73	69	69	69	66	65	65	60	51	67%
40 codex entries Tooling Reasoning	76	75	75	66	66	66	62	55	53	51	64%
tiers											52.42%
36.91%

Data extraction

Extract key details from a given block of text.

Scenario	#1	#2	#3	#4	#5	#6	#7	#8	#9	#10	Total
All valid emails Utility	100	100	100	100	100	100	100	100	100	100	100%
Contextual pronoun Utility	100	100	100	100	100	100	100	100	100	100	100%
Fruits excluding citrus Utility	100	100	100	100	100	100	100	100	100	100	100%
Future event time Reasoning	50	50	50	50	50	50	50	50	50	50	50%
Guess the pet Reasoning	100	100	100	50	50	50	50	50	50	50	65%
Highest-rated movie Reasoning	100	100	100	100	100	50	50	50	50	50	75%
Indirect birth year Reasoning	50	50	50	50	50	50	50	50	50	50	50%
What instrument does Lucy play? Reasoning	100	100	100	50	50	50	50	50	50	50	65%
What's the color of the car? Utility	100	100	100	100	50	50	50	50	50	50	70%
What's the correct time? Reasoning	100	0	0	0	0	0	0	0	0	0	10%
Who's the sister? Reasoning	100	100	100	100	100	100	100	100	100	50	95%
Who's the tallest? Reasoning	100	100	100	100	100	100	100	100	100	100	100%
73.33%

Dialogue tags

Various tasks related to dialogue tags in text.

Scenario	#1	#2	#3	#4	#5	#6	#7	#8	#9	#10	Total
dialogue-200
Write 200 words with 10% dialogue Rule Following Utility	0	0	0	0	0	0	0	0	0	0	0%
Write 200 words with 50% dialogue Rule Following Utility	40	28	21	20	8	3	1	0	0	0	12%
Write 200 words with 90% dialogue Rule Following Utility	50	44	6	4	1	0	0	0	0	0	11%
dialogue-200											7.54%
dialogue-500
Write 500 words with 30% dialogue Rule Following Utility	49	2	2	0	0	0	0	0	0	0	5%
Write 500 words with 50% dialogue Rule Following Utility	50	41	33	21	3	3	0	0	0	0	15%
Write 500 words with 70% dialogue Rule Following Utility	49	19	13	3	0	0	0	0	0	0	8%
dialogue-500											9.63%
Ungrouped
Write unattributed dialogue Rule Following	100	61	14	14	14	14	14	14	14	1	26%
11.02%

Language Comprehension

Does the model understand more than just English?

Scenario	#1	#2	#3	#4	#5	Total
Asking for directions (Dutch) Language	100	100	100	100	0	80%
Asking for directions (German) Language	100	100	100	100	100	100%
Friend got new kittens (German) Language	100	100	100	100	100	100%
Friend got new kittens (Tagalog) Language	100	100	100	100	100	100%
95.00%

Language Writing

Can the model generate text in different languages?

Scenario	#1	#2	#3	#4	#5	Total
Character dialogue (French) in a story Language	100	100	71	0	0	54%
Character dialogue (German) in a story Language	100	88	63	0	0	50%
Character dialogue (Hindi) in a story Language	100	64	63	57	0	57%
Character dialogue (Italian) in a story Language	100	100	100	56	0	71%
Character dialogue (Spanish) in a story Language	100	100	100	67	0	73%
61.10%

N-Length Sentences

Write sentences with exactly N words

Scenario	#1	#2	#3	#4	#5	#6	#7	#8	#9	#10	Total
Write sentences with 5 words each Utility	91	89	89	85	81	68	64	50	23	10	65%
Write sentences with 10 words each Utility	77	74	66	66	59	57	54	52	46	20	57%
Write sentences with 20 words each Utility	41	21	14	5	5	1	1	0	0	0	9%
43.67%

Novel outline

Handle questions about the outline of a novel in various formats

Scenario	#1	#2	#3	#4	#5	#6	#7	#8	#9	#10	Total
outline-count
Count acts Utility	100	100	100	100	100	100	100	100	100	100	100%
Count chapters Utility	100	100	100	100	100	100	100	100	100	100	100%
Count scenes Utility	100	100	0	0	0	0	0	0	0	0	20%
outline-count											73.33%
pov-count
Count point of views for Jack and Olivia Utility	100	100	100	50	50	50	50	50	0	0	55%
Count point of views for Jack Harper Utility	100	100	0	0	0	0	0	0	0	0	20%
Count point of views for Olivia Utility	100	100	0	0	0	0	0	0	0	0	20%
pov-count											31.67%
52.50%

Text Replacement

Tests deterministic text transformations: renaming characters/locations, expanding contractions, tense rewriting, POV shifts, gender swaps, combined transformations, and word avoidance. Scored by checking each expected change independently.

Scenario	#1	#2	#3	#4	#5	#6	#7	Total
Generic Prompt
Avoid said/asked/replied/answered Text Editing	100	100	100	100	100	100	100	100%
Character rename: Elena->Mirabel, Gregor->Aldric Text Editing	100	100	100	100	100	100	100	100%
Combined: 3rd person past → 1st person present Text Editing	99	98	97	96	96	95	93	96%
Expand all contractions Text Editing	100	100	100	100	100	100	100	100%
Location rename: market square, outer ring, bridge, northern mines Text Editing	100	100	100	100	100	100	100	100%
Multi-character gender swap: Priya(F)->Rohan(M), Mara unchanged Text Editing	100	74	74	74	74	74	74	77%
Passive voice → active voice Text Editing Hallucination	94	93	92	86	72	27	3	67%
POV shift: 3rd person to 1st person (Elena's perspective) Text Editing	100	100	100	100	68	66	0	76%
Tense rewriting: past to present Text Editing	96	95	92	89	85	85	82	89%
Generic Prompt								89.56%
Specific Prompt
Avoid said/asked/replied/answered Text Editing	100	100	100	100	100	100	100	100%
Character rename: Elena->Mirabel, Gregor->Aldric Text Editing	100	100	100	100	100	100	100	100%
Combined: 3rd person past → 1st person present Text Editing	99	99	99	99	99	99	99	99%
Expand all contractions Text Editing	100	100	100	100	100	100	100	100%
Location rename: market square, outer ring, bridge, northern mines Text Editing	100	100	100	100	100	100	100	100%
Multi-character gender swap: Priya(F)->Rohan(M), Mara unchanged Text Editing	100	100	100	100	100	100	100	100%
Passive voice → active voice Text Editing Hallucination	98	97	96	96	95	94	93	95%
POV shift: 3rd person to 1st person (Elena's perspective) Text Editing	100	100	100	100	100	100	100	100%
Tense rewriting: past to present Text Editing	100	100	100	100	100	100	100	100%
Specific Prompt								99.43%
94.49%

Tool usage within Novelcrafter

Output messages that are related to tool usage within Novelcrafter

Scenario	#1	#2	#3	#4	#5	#6	#7	#8	#9	#10	Total
Create alternate prose sections Tooling	100	100	100	100	100	100	100	100	100	67	97%
96.67%

Voice/dialogue sheets

Extract dialogue from given text as voice sheets.

Scenario	#1	#2	#3	#4	#5	#6	#7	#8	#9	#10	Total
Multiple speakers Rule Following	0	0	0	0	0	0	0	0	0	0	0%
Simple Rule Following	100	0	0	0	0	0	0	0	0	0	10%
Simple (1-shot) 1-shot Rule Following	100	100	100	100	100	100	100	100	100	100	100%
Simple (5-shot) Few-shot Rule Following	100	100	100	100	100	100	100	100	100	100	100%
Unattributed dialogue Rule Following	0	0	0	0	0	0	0	0	0	0	0%
42.00%

Write N of X

Write exactly N words/sentences/paragraphs...

Scenario	#1	#2	#3	#4	#5	#6	#7	#8	#9	#10	Total
paragraphs
1 paragraph summary Utility	100	100	100	100	100	100	100	100	100	100	100%
3 paragraph summary Utility	100	100	100	100	100	100	100	100	100	100	100%
5 paragraph summary Utility	100	100	100	100	100	100	100	0	0	0	70%
paragraphs											90.00%
sentences
1 sentence summary Utility	100	100	100	100	100	100	100	100	100	100	100%
3 sentence summary Utility	100	100	100	100	100	100	100	100	100	100	100%
10 sentence summary Utility	100	100	100	98	98	98	98	98	77	9	88%
20 sentence summary Utility	100	100	98	98	92	92	77	27	9	2	70%
50 sentence summary Utility	0	0	0	0	0	0	0	0	0	0	0%
sentences											71.48%
words
10 word summary Utility	100	100	100	98	98	77	54	2	0	0	63%
20 word summary Utility	100	98	98	0	0	0	0	0	0	0	30%
50 word summary Utility	0	0	0	0	0	0	0	0	0	0	0%
100 word summary Utility	0	0	0	0	0	0	0	0	0	0	0%
200 word summary Utility	0	0	0	0	0	0	0	0	0	0	0%
words											18.52%
55.39%