Ministral 3B - NC Bench

ministral-3b-2410

Ministral 3B

via mistral

Release Date

Oct 16th, 2024

Context Size

128k

Reasoning

Benchmark Cost

$0.14

Speed

–

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	77	74	72	71	60	71%
Horror: alone in an eerie place at night Creative Writing Hallucination	79	79	75	74	72	76%
Literary fiction: old friends reunite Creative Writing Hallucination	83	79	79	77	75	79%
Mystery: examining a crime scene Creative Writing Hallucination	76	74	74	72	70	73%
Romance: separated couple reunites Creative Writing Hallucination	77	76	75	74	71	75%
Thriller: chase through city streets Creative Writing Hallucination	82	80	78	74	70	77%
Detailed Writing Rules						74.99%
genre
Fantasy: entering an ancient ruin Creative Writing Hallucination	71	70	69	67	66	68%
Horror: alone in an eerie place at night Creative Writing Hallucination	78	78	76	76	69	75%
Literary fiction: old friends reunite Creative Writing Hallucination	80	79	79	79	70	78%
Mystery: examining a crime scene Creative Writing Hallucination	82	76	73	72	71	75%
Romance: separated couple reunites Creative Writing Hallucination	82	78	76	67	65	74%
Thriller: chase through city streets Creative Writing Hallucination	84	79	78	74	72	77%
genre						74.51%
Novelcrafter Default Prompt
Fantasy: entering an ancient ruin Creative Writing Hallucination	77	73	72	69	66	71%
Horror: alone in an eerie place at night Creative Writing Hallucination	79	79	76	75	74	77%
Literary fiction: old friends reunite Creative Writing Hallucination	82	81	78	76	70	77%
Mystery: examining a crime scene Creative Writing Hallucination	78	77	77	74	73	76%
Romance: separated couple reunites Creative Writing Hallucination	82	79	78	74	73	77%
Thriller: chase through city streets Creative Writing Hallucination	82	78	75	74	74	76%
Novelcrafter Default Prompt						75.88%
75.12%

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	92	89	89	88	87	89%
Medium: The Hollow (Inferred) Tooling Reasoning	86	85	85	85	84	85%
Medium: Through the Thornveil (Scattered) Tooling Reasoning	93	89	89	87	83	88%
Short: The Rusty Lantern (Explicit) Tooling Reasoning	93	92	90	90	89	91%
88.15%

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	38	32	29	28	28	27	26	26	5	3	24%
Long text (~1594 words), small codex (11 entries) Hallucination	42	38	31	31	31	31	30	29	29	3	29%
Short text (~524 words), big codex (51 entries) Hallucination	32	31	30	30	30	29	29	28	28	27	29%
Short text (~524 words), small codex (11 entries) Hallucination	50	30	30	29	28	28	28	28	27	8	29%
basic entries											27.87%
detailed entries
Long text (~1594 words), big codex (51 detailed entries) Hallucination	27	27	27	26	26	26	26	2	1	0	19%
Long text (~1594 words), small codex (11 detailed entries) Hallucination	32	31	29	28	27	27	26	26	26	26	28%
Short text (~524 words), big codex (51 detailed entries) Hallucination	42	38	32	29	27	27	26	6	5	5	24%
Short text (~524 words), small codex (11 detailed entries) Hallucination	50	42	33	32	30	29	29	29	29	4	31%
detailed entries											25.18%
26.52%

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	46	44	44	43	43	43	42	41	40	39	43%
Large codex (40 entries), short passage (165 words) Tooling Reasoning	80	79	78	76	71	69	67	60	52	50	68%
Small codex (7 entries), long passage (734 words) Tooling Reasoning	80	65	62	62	62	62	60	53	52	48	61%
Small codex (7 entries), short passage (165 words) Tooling Reasoning	84	80	79	76	75	75	75	71	70	66	75%
matrix											61.69%
tiers
5 codex entries Tooling Reasoning	83	78	76	67	57	54	50	50	44	42	60%
10 codex entries Tooling Reasoning	72	66	61	60	49	49	40	38	33	33	50%
20 codex entries Tooling Reasoning	65	62	59	58	58	53	53	50	47	39	54%
40 codex entries Tooling Reasoning	83	80	72	66	60	60	60	55	55	55	65%
tiers											57.23%
59.46%

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	0	0	80%
Contextual pronoun Utility	100	100	100	100	100	100	100	100	0	0	80%
Fruits excluding citrus Utility	100	100	100	100	100	100	100	100	100	0	90%
Future event time Reasoning	50	50	50	50	50	50	50	50	50	50	50%
Guess the pet Reasoning	100	100	100	100	100	100	100	100	100	100	100%
Highest-rated movie Reasoning	100	100	100	100	100	100	100	100	100	100	100%
Indirect birth year Reasoning	100	50	50	50	50	50	50	0	0	0	40%
What instrument does Lucy play? Reasoning	100	50	0	0	0	0	0	0	0	0	15%
What's the color of the car? Utility	100	100	100	100	100	100	100	100	100	100	100%
What's the correct time? Reasoning	100	100	0	0	0	0	0	0	0	0	20%
Who's the sister? Reasoning	100	100	100	100	100	100	100	100	100	100	100%
Who's the tallest? Reasoning	100	100	100	100	100	100	100	100	100	100	100%
72.92%

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	47	26	22	10	10	0	0	0	0	0	11%
Write 200 words with 50% dialogue Rule Following Utility	76	50	48	47	46	34	18	0	0	0	32%
Write 200 words with 90% dialogue Rule Following Utility	84	50	46	30	30	21	18	15	0	0	29%
dialogue-200											24.30%
dialogue-500
Write 500 words with 30% dialogue Rule Following Utility	65	3	0	0	0	0	0	0	0	0	7%
Write 500 words with 50% dialogue Rule Following Utility	48	48	16	3	3	1	0	0	0	0	12%
Write 500 words with 70% dialogue Rule Following Utility	80	43	24	18	15	0	0	0	0	0	18%
dialogue-500											12.28%
Ungrouped
Write unattributed dialogue Rule Following	100	100	100	61	14	14	14	1	0	0	40%
21.42%

Language Comprehension

Does the model understand more than just English?

Scenario	#1	#2	#3	#4	#5	Total
Asking for directions (Dutch) Language	100	100	0	0	0	40%
Asking for directions (German) Language	100	0	0	0	0	20%
Friend got new kittens (German) Language	100	0	0	0	0	20%
Friend got new kittens (Tagalog) Language	100	0	0	0	0	20%
25.00%

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	50	0	64%
Character dialogue (German) in a story Language	67	64	58	58	0	49%
Character dialogue (Hindi) in a story Language	60	53	50	50	50	53%
Character dialogue (Italian) in a story Language	100	86	64	57	56	73%
Character dialogue (Spanish) in a story Language	86	57	55	50	46	59%
59.49%

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	71	64	64	64	52	52	40	37	20	0	46%
Write sentences with 10 words each Utility	39	32	31	30	27	26	26	21	7	6	24%
Write sentences with 20 words each Utility	23	19	15	11	4	0	0	0	0	0	7%
26.07%

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	0	0	80%
Count scenes Utility	0	0	0	0	0	0	0	0	0	0	0%
outline-count											60.00%
pov-count
Count point of views for Jack and Olivia Utility	100	100	100	50	50	50	0	0	0	0	45%
Count point of views for Jack Harper Utility	100	100	100	0	0	0	0	0	0	0	30%
Count point of views for Olivia Utility	100	100	100	100	0	0	0	0	0	0	40%
pov-count											38.33%
49.17%

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	69	69	61	58	56	53	44	59%
Character rename: Elena->Mirabel, Gregor->Aldric Text Editing	100	100	95	95	95	95	77	94%
Combined: 3rd person past → 1st person present Text Editing	96	96	90	81	76	69	65	82%
Expand all contractions Text Editing	59	59	49	42	42	39	39	47%
Location rename: market square, outer ring, bridge, northern mines Text Editing	96	76	71	70	67	61	61	72%
Multi-character gender swap: Priya(F)->Rohan(M), Mara unchanged Text Editing	88	88	88	88	85	85	85	87%
Passive voice → active voice Text Editing Hallucination	65	62	62	61	55	52	51	58%
POV shift: 3rd person to 1st person (Elena's perspective) Text Editing	94	79	61	60	60	56	52	66%
Tense rewriting: past to present Text Editing	97	78	78	72	64	61	54	72%
Generic Prompt								70.72%
Specific Prompt
Avoid said/asked/replied/answered Text Editing	64	64	64	61	61	50	47	59%
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	92	92	92	92	92	92	91	92%
Expand all contractions Text Editing	97	97	97	97	96	96	96	97%
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	93	99%
Passive voice → active voice Text Editing Hallucination	79	78	78	78	76	76	64	75%
POV shift: 3rd person to 1st person (Elena's perspective) Text Editing	100	100	100	100	98	98	96	99%
Tense rewriting: past to present Text Editing	100	100	100	100	99	99	95	99%
Specific Prompt								91.11%
80.92%

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	0	0	80%
80.00%

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	100	0	0	0	0	0	0	0	0	20%
Simple (1-shot) 1-shot Rule Following	100	100	0	0	0	0	0	0	0	0	20%
Simple (5-shot) Few-shot Rule Following	100	0	0	0	0	0	0	0	0	0	10%
Unattributed dialogue Rule Following	0	0	0	0	0	0	0	0	0	0	0%
10.00%

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	0	0	0	0	0	0	0	0	0	0	0%
5 paragraph summary Utility	0	0	0	0	0	0	0	0	0	0	0%
paragraphs											33.33%
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	98	100%
10 sentence summary Utility	100	100	98	98	92	77	77	27	0	0	67%
20 sentence summary Utility	100	98	77	27	27	9	0	0	0	0	34%
50 sentence summary Utility	100	0	0	0	0	0	0	0	0	0	10%
sentences											62.17%
words
10 word summary Utility	98	98	92	27	27	2	0	0	0	0	35%
20 word summary Utility	92	92	54	0	0	0	0	0	0	0	24%
50 word summary Utility	9	9	2	2	0	0	0	0	0	0	2%
100 word summary Utility	98	0	0	0	0	0	0	0	0	0	10%
200 word summary Utility	0	0	0	0	0	0	0	0	0	0	0%
words											14.07%
37.01%