Compare / Claude Fable 5.2 vs Kimi K3.1 Pro
Claude Fable 5.2 vs Kimi K3.1 Pro
Which is cheaper, which is more capable, and which does better at each kind of work, from 83 results on 5 sources that measured both models.
AnthropicClaude Fable 5.2
Moonshot AIKimi K3.1 Pro
Claude Fable 5.2 wins 4021 too close to callKimi K3.1 Pro wins 22
Of the 83 results both models have, the better value on each. Results whose reported ranges overlap are too close to call.
Results as of 10 November 2026 (catalogue release 2026-11-10-8342a7178929). See also Anthropic vs Moonshot AI, every model from both labs.
Claude Fable 5.2 or Kimi K3.1 Pro? The short answer
Pick Claude Fable 5.2 if quality matters most: it scores higher (Spring Prompt overall 90 against 68). Pick Kimi K3.1 Pro if cost matters more: it costs 48% less per product listing on CatalogBench.
- Spring Prompt overall: Claude Fable 5.2 90, Kimi K3.1 Pro 68 (out of 100; how the score works).
- Kimi K3.1 Pro is cheaper to run: $0.045 against $0.088 per product listing on CatalogBench (48% less).
- Claude Fable 5.2 is newer: released 9 Nov 2026, against 23 Oct 2026 for Kimi K3.1 Pro.
- On our own benchmarks, on the results both models have: Claude Fable 5.2 does better for user surveys and marketing planning.
Head to head
90
Spring Prompt overall
68
$16.00
Price per million tokens, blendedlower is better
$3.81
1,000,000
Context window, tokens
1,048,576
Every fact side by side
| Claude Fable 5.2 | Kimi K3.1 Pro | |
|---|---|---|
| Price per million tokens | $8.00 in · $40.00 out | $1.75 in · $10.00 out |
| Blended (3 in : 1 out) | $16.00 | $3.81 |
| Cheapest host, blended | $16.00 (Amazon Bedrock) | $2.38 (Relace, FP4) |
| Cost per product listing (CatalogBench) | $0.088 | $0.045 |
| Cached input, per million | $0.20 cached input | – |
| Intelligence Index | – | – |
| Spring Prompt overall | 90 | 68 |
| Output speed, tokens a second | – | – |
| Context window | 1,000,000 tokens | 1,048,576 tokens |
| Released | 9 Nov 2026 (newer) | 23 Oct 2026 |
| Weights | Closed | Open |
| Developer based in | the US | China |
Prices are the developer's list price where we have read it, otherwise the typical host on OpenRouter; see every model's price. Intelligence and speed from Artificial Analysis.
Which is better for what
Each model's rank among the models on sale today, for each kind of work, from our own benchmarks and licensed sources. Further right is better; the better rank is in green.
Product listingsTurning a sparse product feed and photos into listings that can go live
5 of 29
not measured
Decks from an analysisTurning a finished analysis into a deck you could present as it is
3 of 29
12 of 29
User surveysPlanning a user survey and reading its results without being misled
1 of 20
18 of 20
Marketing planningPlanning a year of ad spend without overspending
1 of 29
4 of 29
Agents and tool useMulti-step tasks with tools: support desks, coding agents, function calls
not measured
1 of 159
Reasoning and knowledgeHard questions across science, maths and general knowledge
not measured
9 of 187
Speed under pressureGood decisions against a real clock (fast chess)
7 of 32
24 of 32
Their slides, side by side
The same brief for both: Manchester store investment review for Hearthside Coffee, an invented company. First slide of each deck from DeckBench.
From our research
- Claude Fable 5.2 review: first on our survey and ad budget tests, and one of the dearest models to run 10 Nov 2026
- Kimi K3.1 Pro first look: the best open-weight model on our board, and very slow 5 Nov 2026
- GPT-6.1 Terra review: close to the top of our business tests at half the price of Sol 5 Nov 2026
Every shared result
Each model's best result on every measure both have, grouped by benchmark, with the setting that produced it. Under each measure, every model that benchmark has measured is a grey tick, best to the right, so you can see where the two sit in the field. The headline measures of each benchmark come first; the rest are under "Show all". Green marks the better value; where a source reports ranges that overlap, neither is marked.
| Measure | Claude Fable 5.2 | Kimi K3.1 Pro |
|---|---|---|
| BulletBench measured by Spring Prompt · 3 results | ||
| Cost per game · Bullet 60sUS dollars, lower is better 37 · 28 of 38 measured | $0.0800low reasoning | $0.0175low reasoning |
| Ladder Elo · Bullet 60sladder Elo, higher is better 15 · 31 of 38 measured | 491 | 147low reasoning |
| Games lost on time · Bullet 60s% of games, lower is better 27 · 34 of 38 measured | 62.5% | 87.5%low reasoning |
| CatalogBench measured by Spring Prompt · 3 results | ||
| Cost per productUS dollars, lower is better 29 · 27 of 30 measured | $0.0875 | $0.0454 |
| Channel compliance% of products, higher is better 1 · 1 of 30 measured | 100.0% | 100.0% |
| Channel rules broken% of products, lower is better 1 · 1 of 30 measured | 0.0% | 0.0% |
| DeckBench measured by Spring Prompt · 3 results | ||
| Cost per deckUS dollars, lower is better 30 · 27 of 30 measured | $1.40 | $0.54 |
| Deck ratingrating, higher is betterThe rating weighs each win by how strong the other deck was; the win rate counts wins against whichever decks a model happened to meet, so the two can disagree. 3 · 12 of 30 measured | 1,419 | 1,238 |
| Accurate decks% of tasks, higher is better 5 · 9 of 30 measured | 83.3% | 66.7% |
| ROASBench measured by Spring Prompt · 3 results | ||
| Overall scorescore out of 100, higher is better 1 · 4 of 30 measured | 61.3 | 55.1 |
| Audience scorescore out of 100, higher is better 4 · 8 of 30 measured | 68.5 | 67.6 |
| Business scorescore out of 100, higher is better 1 · 3 of 30 measured | 72.3 | 58.2 |
| SurveyBench measured by Spring Prompt · 3 results | ||
| Analysis ratingrating, higher is betterThe rating weighs each win by how strong the other analysis was; the win rate counts wins against whichever analyses a model happened to meet, so the two can disagree. 1 · 19 of 21 measured | 1,664 | 608 |
| Cost per taskUS dollars, lower is better 21 · 20 of 21 measured | $1.51 | $0.74 |
| Analysis scorepoints of 100, higher is better 1 · 20 of 21 measured | 98.0 | 14.5 |
Show all 83 results (68 more)
| Measure | Claude Fable 5.2 | Kimi K3.1 Pro |
|---|---|---|
| BulletBench measured by Spring Prompt · 2 results | ||
| Invalid moves · Bullet 60s% of moves, lower is better 1 · 1 of 38 measured | 0.0% | 0.0%low reasoning |
| Median move time · Bullet 60smilliseconds, lower is better 33 · 38 of 38 measured | 3.0 slow reasoning | 5.0 slow reasoning |
| CatalogBench measured by Spring Prompt · 27 results | ||
| Channel rules broken · sales brief% of products, lower is better 1 · 1 of 30 measured | 0.0% | 0.0% |
| Claims to check% of products, lower is better 1 · 1 of 28 measured | 0.0% | 0.0% |
| Claims to check · sales brief% of products, lower is better 1 · 1 of 28 measured | 0.0% | 0.0% |
| Conflicts caught% of conflicts, higher is better 1 · 1 of 30 measured | 100.0% | 100.0% |
| Content quality% of checks, higher is better 2 · 1 of 30 measured | 98.6% | 98.9% |
| Content quality · sales brief% of checks, higher is better 3 · 24 of 30 measured | 98.6% | 95.3% |
| Decision accuracy% of decisions, higher is better 2 · 1 of 30 measured | 98.7% | 98.9% |
| Failed outputs% of products, lower is better 1 · 1 of 30 measured | 0.0% | 0.0% |
| Failed outputs · sales brief% of products, lower is better 1 · 28 of 30 measured | 0.0% | 3.6% |
| Field accuracy% of missing fields, higher is better 1 · 3 of 30 measured | 98.1% | 95.6% |
| Invented values% of filled values, lower is better 19 · 21 of 30 measured | 1.9% | 2.1% |
| Missing UK information% of products, lower is better 1 · 1 of 30 measured | 0.0% | 0.0% |
| Missing UK information · sales brief% of products, lower is better 1 · 1 of 30 measured | 0.0% | 0.0% |
| Not findable% of products, lower is better 5 · 17 of 30 measured | 1.8% | 5.4% |
| Not findable · sales brief% of products, lower is better 5 · 14 of 30 measured | 1.8% | 3.6% |
| Output tokens per producttokens, lower is better 3 · 8 of 8 measured | 1,442 | 3,996 |
| Publish-ready listings% of products, higher is better 1 · 2 of 30 measured | 81.0% | 78.6% |
| Publish-ready listings · sales brief% of products, higher is better 8 · 2 of 30 measured | 66.1% | 75.0% |
| Time per productseconds, lower is better 17 · 30 of 30 measured | 17.5 | 114 |
| Unsupported claimsclaims per product, lower is better 7 · 1 of 30 measured | 0.23 | 0.00 |
| Unsupported claims% of products, lower is better 6 · 1 of 30 measured | 6.0% | 0.0% |
| Unsupported claims · sales brief% of products, lower is better 10 · 1 of 30 measured | 23.8% | 1.8% |
| Unsupported claims · sales briefclaims per product, lower is better 8 · 5 of 30 measured | 0.46 | 0.21 |
| Wrong attributes% of products, lower is better 1 · 13 of 30 measured | 9.5% | 16.1% |
| Wrong attributes · sales brief% of products, lower is better 3 · 5 of 30 measured | 13.1% | 14.3% |
| Wrong category or variant% of products, lower is better 5 · 5 of 30 measured | 1.8% | 1.8% |
| Wrong category or variant · sales brief% of products, lower is better 6 · 27 of 30 measured | 1.8% | 5.4% |
| DeckBench measured by Spring Prompt · 15 results | ||
| Caveat dropped% of tasks, lower is better 14 · 14 of 30 measured | 16.7% | 16.7% |
| Clean layout% of tasks, higher is better 3 · 27 of 30 measured | 83.3% | 0.0% |
| Design quality% of the maximum, higher is better 8 · 3 of 30 measured | 77.8% | 83.9% |
| Draft figure quoted% of tasks, lower is better 12 · 1 of 30 measured | 16.7% | 0.0% |
| Findings missing% of tasks, lower is better 20 · 28 of 30 measured | 16.7% | 33.3% |
| Head-to-head win rate% of comparisons, higher is betterThe rating weighs each win by how strong the other deck was; the win rate counts wins against whichever decks a model happened to meet, so the two can disagree. 6 · 10 of 30 measured | 78.0% | 71.1% |
| Layout defects% of tasks, lower is better 3 · 27 of 30 measured | 16.7% | 100.0% |
| Misleading metric used% of tasks, lower is better 26 · 1 of 30 measured | 16.7% | 0.0% |
| Output tokens per decktokens, lower is better 7 · 9 of 9 measured | 33,285 | 51,739 |
| Presentable decks% of tasks, higher is better 1 · 14 of 30 measured | 83.3% | 0.0% |
| Recommendation late or wrong% of tasks, lower is better 12 · 1 of 30 measured | 16.7% | 0.0% |
| Slides needing work% of tasks, lower is better 3 · 1 of 30 measured | 16.7% | 0.0% |
| Time per deckseconds, lower is better 24 · 30 of 30 measured | 285 | 1,288 |
| Unsupported claims% of tasks, lower is better 15 · 1 of 30 measured | 16.7% | 0.0% |
| Unsupported numbers% of tasks, lower is better 24 · 1 of 30 measured | 16.7% | 0.0% |
| ROASBench measured by Spring Prompt · 6 results | ||
| Consistency scorescore out of 100, higher is better 12 · 6 of 30 measured | 31.6 | 36.3 |
| Contribution profitsimulated US dollars, higher is better 1 · 4 of 30 measured | $1,032,894 | $858,023 |
| Cost of a runUS dollars, lower is better 29 · 28 of 30 measured | $1.76 | $1.60 |
| Months over budgetmonths of 12, lower is better 1 · 1 of 30 measured | 0 | 0 |
| Planning scorescore out of 100, higher is better 27 · 4 of 30 measured | 54.2 | 55.0 |
| Return on ad spendprofit per $1 spent, higher is better 1 · 2 of 30 measured | $0.96 | $0.72 |
| SurveyBench measured by Spring Prompt · 18 results | ||
| Head-to-head win rate% of comparisons, higher is betterThe rating weighs each win by how strong the other analysis was; the win rate counts wins against whichever analyses a model happened to meet, so the two can disagree. 1 · 15 of 21 measured | 86.7% | 36.1% |
| Leading or double-barrelled questions% of tasks, lower is better 7 · 7 of 21 measured | 22.2% | 22.2% |
| Mistaken next steps% of tasks, lower is better 1 · 19 of 21 measured | 0.0% | 77.8% |
| Numbers right% of answers, higher is better 1 · 1 of 21 measured | 100.0% | 100.0% |
| Objectives unanswerable% of tasks, lower is better 1 · 1 of 21 measured | 0.0% | 0.0% |
| Output tokens per tasktokens, lower is better 6 · 11 of 11 measured | 33,902 | 70,935 |
| Precision ignored% of tasks, lower is better 1 · 1 of 21 measured | 0.0% | 0.0% |
| Research scorepoints of 100, higher is better 1 · 20 of 21 measured | 97.6 | 39.2 |
| Sound research% of tasks, higher is better 3 · 12 of 21 measured | 44.4% | 0.0% |
| Stakeholder's view presumed% of tasks, lower is better 1 · 1 of 21 measured | 0.0% | 0.0% |
| Survey plan scorepoints of 100, higher is better 4 · 3 of 21 measured | 96.9 | 97.0 |
| Survey structure% of tasks, lower is better 3 · 3 of 21 measured | 11.1% | 11.1% |
| Time per taskseconds, lower is better 17 · 21 of 21 measured | 337 | 1,956 |
| Traps handled% of traps, higher is better 2 · 10 of 21 measured | 98.6% | 93.3% |
| Traps missed% of tasks, lower is better 2 · 13 of 21 measured | 11.1% | 88.9% |
| Unsupported findings% of tasks, lower is better 3 · 12 of 21 measured | 11.1% | 77.8% |
| Wrong numbers% of tasks, lower is better 1 · 17 of 21 measured | 0.0% | 77.8% |
| Wrong recommendation% of tasks, lower is better 1 · 21 of 21 measured | 0.0% | 100.0% |
Compare them with other models
Quick answers
Is Claude Fable 5.2 better than Kimi K3.1 Pro?
Pick Claude Fable 5.2 if quality matters most: it scores higher (Spring Prompt overall 90 against 68). Pick Kimi K3.1 Pro if cost matters more: it costs 48% less per product listing on CatalogBench.
Which is cheaper, Claude Fable 5.2 or Kimi K3.1 Pro?
Kimi K3.1 Pro is cheaper to run: $0.045 against $0.088 per product listing on CatalogBench (48% less).
Which is better for decks from an analysis, Claude Fable 5.2 or Kimi K3.1 Pro?
Claude Fable 5.2 ranks 3 of 29 models on sale for turning a finished analysis into a deck you could present as it is, against 12 for Kimi K3.1 Pro.
Which is better for user surveys, Claude Fable 5.2 or Kimi K3.1 Pro?
Claude Fable 5.2 ranks 1 of 20 models on sale for planning a user survey and reading its results without being misled, against 18 for Kimi K3.1 Pro.
Which is better for marketing planning, Claude Fable 5.2 or Kimi K3.1 Pro?
Claude Fable 5.2 ranks 1 of 29 models on sale for planning a year of ad spend without overspending, against 4 for Kimi K3.1 Pro.
Which is better for speed under pressure, Claude Fable 5.2 or Kimi K3.1 Pro?
Claude Fable 5.2 ranks 7 of 32 models on sale for good decisions against a real clock (fast chess), against 24 for Kimi K3.1 Pro.
Which has the bigger context window, Claude Fable 5.2 or Kimi K3.1 Pro?
Kimi K3.1 Pro: 1,048,576 tokens, against 1,000,000 for Claude Fable 5.2.
Run Claude Fable 5.2 and Kimi K3.1 Pro on your own prompt.
Benchmarks aren't your task. Test both side by side on your prompt, with every answer checked against your rules and the cost of each one.