Claude Fable 5
VS
MiniMax M2.7
Claude Fable 5 vs MiniMax M2.7: which wins at real work?
22 task areas · same graded test runs · rank comparison only, so 0–100 and Elo collections never mix raw scores.
Claude Fable 5 wins 20 of 22 task areas we tested; MiniMax M2.7 takes 2. MiniMax M2.7 costs 66.7× less per token ($0.9 vs $60 per 1M).
MiniMax M2.7 costs 66.7× less per token ($0.9 vs $60 per 1M).
Task by task
| Task area | Claude Fable 5 | MiniMax M2.7 | Winner |
|---|---|---|---|
| Chef / Home Cooking |
#3
/ 126
Strong
|
#124
/ 126
Weak
|
Claude Fable 5 |
| AI Strategy |
#1
/ 126
Strong
|
#119
/ 126
Usable
|
Claude Fable 5 |
| Creative & Comedy | #1 / 110 | #108 / 110 | Claude Fable 5 |
| Customer Support |
#1
/ 113
Excellent
|
#108
/ 113
Needs editing
|
Claude Fable 5 |
| Sales |
#5
/ 110
Strong
|
#108
/ 110
Weak
|
Claude Fable 5 |
| Content & Brand |
#1
/ 124
Excellent
|
#96
/ 124
Usable
|
Claude Fable 5 |
| Coding |
#11
/ 115
Excellent
|
#105
/ 115
Usable
|
Claude Fable 5 |
| Research & Competitive Analysis |
#3
/ 110
Excellent
|
#96
/ 110
Needs editing
|
Claude Fable 5 |
| Training & Education |
#16
/ 110
Excellent
|
#106
/ 110
Usable
|
Claude Fable 5 |
| Data & Analytics |
#3
/ 110
Excellent
|
#92
/ 110
Excellent
|
Claude Fable 5 |
| Executive Assistant |
#22
/ 112
Strong
|
#107
/ 112
Usable
|
Claude Fable 5 |
| Knowledge & Docs |
#27
/ 110
Strong
|
#108
/ 110
Weak
|
Claude Fable 5 |
| Legal & HR |
#1
/ 110
Excellent
|
#81
/ 110
Strong
|
Claude Fable 5 |
| Summarization & Meeting Notes |
#24
/ 110
Excellent
|
#104
/ 110
Strong
|
Claude Fable 5 |
| Landing Pages |
#1
/ 72
Strong
|
#68
/ 72
Needs editing
|
Claude Fable 5 |
| RAG, Safety & Grounding |
#30
/ 113
Excellent
|
#94
/ 113
Strong
|
Claude Fable 5 |
| Translation & Localization |
#18
/ 110
Excellent
|
#82
/ 110
Excellent
|
Claude Fable 5 |
| Investor & Pitch |
#3
/ 66
Strong
|
#66
/ 66
Weak
|
Claude Fable 5 |
| Presentations & Decks |
#20
/ 110
Excellent
|
#83
/ 110
Strong
|
Claude Fable 5 |
| Structured Output |
#68
/ 113
Strong
|
#33
/ 113
Excellent
|
MiniMax M2.7 |
| Frontend & Landing Pages |
#58
/ 109
Needs editing
|
#89
/ 109
Weak
|
Claude Fable 5 |
| Product & Project Management |
#76
/ 110
Strong
|
#64
/ 110
Strong
|
MiniMax M2.7 |
Rank = position among every model config we tested in that task area (lower is better). Sorted by biggest gap first.
Same task, both models — judged
Both models answered the same test case; an independent judge graded each. Quotes are the judge's actual rationale.
Chef / Home Cooking
Pasta sauce without cream (Substitution Test)
“The response is expert-level, providing highly accurate culinary techniques (tempering, off-heat mixing) to prevent yogurt from splitting while perfectly utilizing the available ingredients. The flavour tradeoffs are insightful, realistic, and the quantities are perfectly scaled for two.”
“The model completely failed the task by outputting its internal planning process, which devolved into an infinite repetition loop of 'We also need to mention...' before cutting off. No usable recipe or advice was provided.”
Thai dinner with satay, stir fry, rice (Meal Timing Test)
“The model provides an expert-level, highly realistic, and perfectly sequenced cooking timeline that strictly adheres to all constraints.”
“The model completely failed the task by getting stuck in a repetitive loop during its planning phase. It never actually generated the requested timeline, make-ahead steps, or plating plan, resulting in a malformed and useless output.”
Creative & Comedy
Roast generic SaaS landing pages (Funny on Command)
“Ranked by Bradley-Terry fit over stored pairwise judgments.”
“Ranked by Bradley-Terry fit over stored pairwise judgments.”
One-liner: the 'quick sync' meeting (Funny on Command)
“Ranked by Bradley-Terry fit over stored pairwise judgments.”
“Ranked by Bradley-Terry fit over stored pairwise judgments.”
Content & Brand
B2B SaaS launch announcement (Generic Copy Index)
“The response perfectly executes the prompt, adhering to all constraints including the strict word count. It uses highly realistic, concrete details and adopts a perfectly restrained, professional tone suited for finance leaders. It avoids all generic AI filler and is difficult to improve.”
“The response failed the word count constraint, included unrequested meta-text, and relied on lazy placeholders instead of providing specific details.”
Frequently asked
Is Claude Fable 5 better than MiniMax M2.7?
Across 22 task areas we benchmarked, Claude Fable 5 ranks higher in 20 and MiniMax M2.7 in 2.
Which is cheaper, Claude Fable 5 or MiniMax M2.7?
MiniMax M2.7 costs 66.7× less per token ($0.9 vs $60 per 1M).
What is Claude Fable 5 better at?
Claude Fable 5 out-ranks MiniMax M2.7 at Chef / Home Cooking, AI Strategy, Creative & Comedy.
What is MiniMax M2.7 better at?
MiniMax M2.7 out-ranks Claude Fable 5 at Structured Output, Product & Project Management.
More comparisons
This page is Spring Prompt, running in public
We just did this for every model. Do it for your prompt.
The rankings above come from running real tasks through real models and scoring every output. Spring Prompt is that same engine — pointed at your prompt, your test cases, and your definition of good.
- Generate test cases from your prompt — no eval set required to start.
- Compare models side by side with quality, cost and latency in one matrix.
- Optimise the winner until the scores say it's ready to ship.
Prompt × model results
12 test cases · 3 evals