Long documents
Prompts that approach the documented context limit
Kimi K3.1 Pro
Kimi K3.1 Pro has the larger documented context window.
Confidence: documented
Claude Fable 5.2 vs Kimi K3.1 Pro
Updated November 9, 2026. Rank says Claude Fable 5.2 is ahead. Price, access, and your workload can each overturn that. Public scores include evidence status and uncertainty.
Share or export
Decision reading
Claude Fable 5.2 has the higher public point estimate, 91.19 versus 74.61. Their conditional score ranges overlap. These ranges do not establish rank confidence. 8 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.
- Shared results
- 8
- Claude Fable 5.2 only
- 22
- Kimi K3.1 Pro only
- 39
- Like-for-like categories
- 2 / 8
Estimated: Claude Fable 5.2 and Kimi K3.1 Pro. Conditional ranges do not establish rank confidence.How the comparison works
Which one for your work
Recommendations appear only when a shared evidence basis or an explicit operating constraint supports the call. Secondary and unsupported use cases stay disclosed below the initial list.
Chat turn cost
1K fresh input + 500 output tokens
Kimi K3.1 Pro
Kimi K3.1 Pro has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-ratesCache-heavy agent loop cost
200K cached + 20K fresh input + 10K output tokens
Kimi K3.1 Pro
Kimi K3.1 Pro has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-rates
Show secondary and unsupported calls
Repository review cost
50K fresh input + 3K output tokens
Kimi K3.1 Pro
Kimi K3.1 Pro has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-ratesCoding work
Code generation, repair, and software-engineering tasks
Not enough matched evidence
Kimi K3.1 Pro is not ranked on the public lane for coding, so no winner is named for coding.
Confidence: limitedAgentic work
Tool use, computer use, and multi-step task completion
Not enough matched evidence
Claude Fable 5.2 is scored on Estimated evidence for agentic, so the reading is directional rather than like-for-like.
Confidence: limited
Which one for a specific job
Choose a job from the LLM Selector's task catalog to see the category row it rests on, under the same basis rules as the table below. A directional row stays directional; choosing a job never creates a winner.
The same task catalog as the LLM Selector. Each job names the evidence surface it rests on; nothing here adds a new score.
86.9Claude Fable 5.2—Kimi K3.1 Pro
Not comparable · BenchAlign v5.8
The coding row is not comparable on the public lane: at least one model is not measured or not ranked there.
Coding scores combine specific tasks and setups. Match the editor, harness, and effort to your workflow.
Same basis rules as the category table below
What is actually comparable
Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.
22 Claude Fable 5.2 only
8 Shared
39 Kimi K3.1 Pro only
2 categories rest on Estimated evidence or different evidence sets. Those rows remain visible for coverage context but do not name a winner.
Shape of the matched evidence
Only shared public evidence is shown. Sparse evidence stays a ruled list rather than being closed into a radar shape.
Too few matched category axes support a radar. The ruled list below shows only shared benchmark results; positions use each benchmark’s normalized display scale when available.
HLE w/o toolsKnowledge
Normalized gap 22.1Claude Fable 5.2:73.2%Kimi K3.1 Pro:51.1%OfficeQA ProMultimodal
Normalized gap 5.5Claude Fable 5.2:72.9%Kimi K3.1 Pro:67.4%HLEKnowledge
Normalized gap 4.2Claude Fable 5.2:74.2%Kimi K3.1 Pro:70%
Bars run 0–100 on each benchmark’s normalized display scale
Category results, on a stated basis
Each row shows the public-lane category score for both models: the BenchAlign v5.8 lane for agentic, coding, reasoning, multimodal, and knowledge, and the weighted public rows elsewhere, with the same rank each model page reports. A row is like-for-like only when both scores rest on Supported evidence or the same weighted set. Directional and not-comparable rows remain visible, but they never receive a winner.
Reasoning
Like-for-like- Claude Fable 5.2
- 90.7
- Supported · #1/157
- Kimi K3.1 Pro
- 70.6
- Supported · #16/157
- Basis
- BenchAlign v5.8 lane · 3 vs 0 public rows
- Reading
- Claude Fable 5.2 leads
Knowledge
Like-for-like- Claude Fable 5.2
- 97.6
- Supported · #1/179
- Kimi K3.1 Pro
- 74.5
- Supported · #13/179
- Basis
- BenchAlign v5.8 lane · 3 vs 5 public rows
- Reading
- Claude Fable 5.2 leads
Agentic
Directional only- Claude Fable 5.2
- 90.7
- Estimated · #1/132
- Kimi K3.1 Pro
- 72.7
- Supported · #10/132
- Basis
- BenchAlign v5.8 lane · 8 vs 19 public rows
- Reading
- Directional only
Multimodal
Directional only- Claude Fable 5.2
- 89.8
- Estimated · #1/67
- Kimi K3.1 Pro
- 73.9
- Estimated · #12/67
- Basis
- BenchAlign v5.8 lane · 6 vs 15 public rows
- Reading
- Directional only
Coding
Not comparable- Claude Fable 5.2
- 86.9
- Estimated · #1/158
- Kimi K3.1 Pro
- Not ranked
- Basis
- BenchAlign v5.8 lane · 8 vs 8 public rows
- Reading
- Not comparable
Multilingual
Not comparable- Claude Fable 5.2
- Not ranked
- Kimi K3.1 Pro
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Instruction following
Not comparable- Claude Fable 5.2
- Not ranked
- Kimi K3.1 Pro
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Math
Not comparable- Claude Fable 5.2
- Not ranked
- Kimi K3.1 Pro
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
| Category | Claude Fable 5.2 | Kimi K3.1 Pro | Basis | Reading |
|---|---|---|---|---|
| Reasoning | 90.7Supported · #1/157 | 70.6Supported · #16/157 | Like-for-likeBenchAlign v5.8 lane · 3 vs 0 public rows | Claude Fable 5.2 leads |
| Knowledge | 97.6Supported · #1/179 | 74.5Supported · #13/179 | Like-for-likeBenchAlign v5.8 lane · 3 vs 5 public rows | Claude Fable 5.2 leads |
| Agentic | 90.7Estimated · #1/132 | 72.7Supported · #10/132 | Directional onlyBenchAlign v5.8 lane · 8 vs 19 public rows | Directional only |
| Multimodal | 89.8Estimated · #1/67 | 73.9Estimated · #12/67 | Directional onlyBenchAlign v5.8 lane · 6 vs 15 public rows | Directional only |
| Coding | 86.9Estimated · #1/158 | Not ranked | Not comparableBenchAlign v5.8 lane · 8 vs 8 public rows | Not comparable |
| Multilingual | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Instruction following | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Math | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
Ranks count the models scored in each category’s lane, so the agentic, coding, reasoning, multimodal, and knowledge denominators (BenchAlign v5.8) differ from the provisional-lane categories. Unranked scores sit on the provisional lane’s scale but never name a winner. Like-for-like rows within 0.5 points read as a practical tie.
Supported evidence per lane · bars run 0–100Methodology
What each workload costs
Three fixed token mixes turn per-token rates into comparable decisions. Each scenario states context fit and whether cached input had to fall back to the published list-input rate.
Chat turn
1K fresh input + 500 output tokens
- Claude Fable 5.2
- $0.028
- Fits in one request
- Kimi K3.1 Pro
- $0.007
- Fits in one request
Kimi K3.1 Pro has the lower modeled cost
Costs use the listed standard API rates.
Repository review
50K fresh input + 3K output tokens
- Claude Fable 5.2
- $0.52
- Fits in one request
- Kimi K3.1 Pro
- $0.13
- Fits in one request
Kimi K3.1 Pro has the lower modeled cost
Costs use the listed standard API rates.
Cache-heavy agent loop
200K cached + 20K fresh input + 10K output tokens
- Claude Fable 5.2
- $0.6
- Fits in one request
- Kimi K3.1 Pro
- $0.18
- Fits in one request
Kimi K3.1 Pro has the lower modeled cost
Costs use the listed standard API rates.
Cached input falls back to the list input rate only where a cached rate is unpublished
Specification differences
Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.
SpecificationClaude Fable 5.2Kimi K3.1 Pro
Context window
Maximum documented context; output-token limits may be lower.
Claude Fable 5.2
Kimi K3.1 Pro
1.05M
API model ID
Claude Fable 5.2
claude-fable-5-2
Anthropic Claude Fable 5.2 model documentationKimi K3.1 Pro
Not sourced
Cached-input rate
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
Claude Fable 5.2
$0.2 per 1M cached input tokens
Claude API pricingKimi K3.1 Pro
$0.2 per 1M cached input tokens
Documented inputs
Claude Fable 5.2
Kimi K3.1 Pro
Not sourced
Documented outputs
Claude Fable 5.2
Kimi K3.1 Pro
Not sourced
Provider availability
Claude Fable 5.2
Generally Available · Claude API, Claude products, AWS, Google Cloud, Microsoft Azure
Anthropic Fable 5.2 and Mythos 5.2 launchKimi K3.1 Pro
Not sourced
Reasoning profile
Claude Fable 5.2
Reasoning
Kimi K3.1 Pro
Reasoning
Weight access
Claude Fable 5.2
Proprietary
Kimi K3.1 Pro
Open Weight
License
Claude Fable 5.2
Proprietary
Kimi K3.1 Pro
Open Weight
Release date
Claude Fable 5.2
2026-11-09
Kimi K3.1 Pro
2026-10-23
If you already use one of these models
- Deployment change
- The models list different providers, so authentication, endpoint behavior, limits, and feature support may change.
- Quality signal
- Claude Fable 5.2 has the higher public point estimate, 91.19 versus 74.61. Their conditional score ranges overlap. These ranges do not establish rank confidence.
- Workload cost
- Repository review: $0.52 vs $0.13. Cache-heavy agent loop: $0.6 vs $0.18.
- Context tradeoff
- Kimi K3.1 Pro has the larger documented window (1.05M).
Run the same representative tasks against both endpoints before changing production traffic.
Questions
Which is better, Claude Fable 5.2 or Kimi K3.1 Pro?
Claude Fable 5.2 has the higher public point estimate, 91.19 versus 74.61. Their conditional score ranges overlap. These ranges do not establish rank confidence. The page therefore keeps the decision tied to the specific documented workload.
Which is better for coding, Claude Fable 5.2 or Kimi K3.1 Pro?
Kimi K3.1 Pro is not ranked on the public lane for coding, so no winner is named for coding.
Which is better for agentic tasks, Claude Fable 5.2 or Kimi K3.1 Pro?
Claude Fable 5.2 scores higher for agentic tasks on the public lane, 90.7 to 72.7. Claude Fable 5.2 is scored on Estimated evidence for agentic tasks, so the reading is directional rather than like-for-like. BenchLM does not name a winner for agentic tasks from a directional reading; read the shared benchmark rows directly and test the models on the same task set.
Which costs less, Claude Fable 5.2 or Kimi K3.1 Pro?
For the stated presets, chat costs $0.028 on Claude Fable 5.2 and $0.007 on Kimi K3.1 Pro; repository review costs $0.52 and $0.13; the cache-heavy agent loop costs $0.6 and $0.18. Costs use the listed standard API rates.
Which has the larger context window, Claude Fable 5.2 or Kimi K3.1 Pro?
Kimi K3.1 Pro has the larger documented context window: 1.05M, compared with 1M.
Self-host vs API cost
Estimates at 50,000 req/day · 1000 tokens/req average.
Claude Fable 5.2
API / mo$36,000
Self-host / moNot listed
Break-even—
Proprietary model — self-hosting not applicable.
Kimi K3.1 Pro
API / mo$9,000
Self-host / mo$57,606
Break-even436M/day
Benchmark evidence
The full public result ledger is available for audit without forcing a wide desktop table onto a phone.
Browse raw public benchmark evidence69 rows
Agentic
Terminal-Bench 4.0
Claude Fable 5.267.60%SourceKimi K3.1 Pro65.70%SourceClaude Fable 5.2 leads this result
Terminal-Bench-Science 0.1
Claude Fable 5.274.6%SourceKimi K3.1 Pro38.1%SourceClaude Fable 5.2 leads this result
AutomationBench
Claude Fable 5.251.4%SourceKimi K3.1 Pro—Not directly comparable
OSWorld 2.1 Offline
Claude Fable 5.289.3%SourceKimi K3.1 Pro—Not directly comparable
WANDR
Claude Fable 5.282.1%SourceKimi K3.1 Pro—Not directly comparable
Remote Labor Index
Claude Fable 5.241.3%SourceKimi K3.1 Pro—Not directly comparable
RoboDojo Sim
Claude Fable 5.264.9%SourceKimi K3.1 Pro—Not directly comparable
DRACO
Claude Fable 5.287.7%SourceKimi K3.1 Pro—Not directly comparable
HLE w/ tools
Claude Fable 5.2—Kimi K3.1 Pro70.0%SourceNot directly comparable
BrowseComp
Claude Fable 5.2—Kimi K3.1 Pro92.1%SourceNot directly comparable
BioMysteryBench
Claude Fable 5.2—Kimi K3.1 Pro80.6%SourceNot directly comparable
Agents' Last Exam
Claude Fable 5.2—Kimi K3.1 Pro43.2%SourceNot directly comparable
Agents' Last Exam (CLI)
Claude Fable 5.2—Kimi K3.1 Pro44.2%SourceNot directly comparable
APEX-Agents
Claude Fable 5.2—Kimi K3.1 Pro69.4%SourceNot directly comparable
Frontier Finance
Claude Fable 5.2—Kimi K3.1 Pro58.6%SourceNot directly comparable
HANDBOOK.md
Claude Fable 5.2—Kimi K3.1 Pro44.6%SourceNot directly comparable
OSWorld 2.0
Claude Fable 5.2—Kimi K3.1 Pro79.6%SourceNot directly comparable
Business Operation
Claude Fable 5.2—Kimi K3.1 Pro37.8%SourceNot directly comparable
Financial Modeling
Claude Fable 5.2—Kimi K3.1 Pro57.6%SourceNot directly comparable
Financial Research
Claude Fable 5.2—Kimi K3.1 Pro70.4%SourceNot directly comparable
Kimi Hallu Bench Hard (tools)
Claude Fable 5.2—Kimi K3.1 Pro8.3%SourceNot directly comparable
ExploitGym
Claude Fable 5.2—Kimi K3.1 Pro20.7%SourceNot directly comparable
MazeBench
Claude Fable 5.2—Kimi K3.1 Pro2%SourceNot directly comparable
OSWorld 2.1
Claude Fable 5.2—Kimi K3.1 Pro79.6%SourceNot directly comparable
CWE-bench
Claude Fable 5.2—Kimi K3.1 Pro45.3%SourceNot directly comparable
Coding
SWE-bench Pro
Claude Fable 5.293%SourceKimi K3.1 Pro—Not directly comparable
SWE Multilingual
Claude Fable 5.296.6%SourceKimi K3.1 Pro—Not directly comparable
SWE Multimodal
Claude Fable 5.274.1%SourceKimi K3.1 Pro—Not directly comparable
DeepSWE
Claude Fable 5.274.3%SourceKimi K3.1 Pro75.5%SourceKimi K3.1 Pro leads this result
FrontierSWE v2
Claude Fable 5.269.9%SourceKimi K3.1 Pro—Not directly comparable
ProgramBench
Claude Fable 5.285.6%SourceKimi K3.1 Pro—Not directly comparable
FrontierCode 2.0
Shared sourceClaude Fable 5.249.1%Kimi K3.1 Pro19.6%Claude Fable 5.2 leads this result
CursorBench 4.0
Claude Fable 5.260.0%SourceKimi K3.1 Pro—Not directly comparable
ProgramBench (binary)
Claude Fable 5.2—Kimi K3.1 Pro31.5%SourceNot directly comparable
Kimi Code Bench v2
Claude Fable 5.2—Kimi K3.1 Pro77.1%SourceNot directly comparable
Kimi Code Bench 3.0 Long-horizon
Claude Fable 5.2—Kimi K3.1 Pro61.2%SourceNot directly comparable
sweMarathon
Claude Fable 5.2—Kimi K3.1 Pro51.7%SourceNot directly comparable
PostTrainBench v1.2
Claude Fable 5.2—Kimi K3.1 Pro36.5%SourceNot directly comparable
Bug Hunt Bench
Claude Fable 5.2—Kimi K3.1 Pro36.0 fixesSourceNot directly comparable
Reasoning
Multimodal
Chartography (no tools)
Claude Fable 5.273.4%SourceKimi K3.1 Pro—Not directly comparable
Chartography (tools)
Claude Fable 5.291.6%SourceKimi K3.1 Pro90.2%SourceClaude Fable 5.2 leads this result
OfficeQA
Claude Fable 5.283.4%SourceKimi K3.1 Pro—Not directly comparable
OfficeQA Pro
Claude Fable 5.272.9%SourceKimi K3.1 Pro67.4%SourceClaude Fable 5.2 leads this result
BenchCAD 2.1
Claude Fable 5.275.4%SourceKimi K3.1 Pro—Not directly comparable
BenchCAD 2.1 (tools)
Claude Fable 5.283.5%SourceKimi K3.1 Pro—Not directly comparable
OfficeQA Pro v2
Claude Fable 5.2—Kimi K3.1 Pro64.4%SourceNot directly comparable
BenchCAD Vision2Code (no tools)
Claude Fable 5.2—Kimi K3.1 Pro0.728SourceNot directly comparable
BenchCAD Vision2Code (tools)
Claude Fable 5.2—Kimi K3.1 Pro0.927SourceNot directly comparable
MMMU-Pro
Claude Fable 5.2—Kimi K3.1 Pro84.1%SourceNot directly comparable
MMMU-Pro w/ Python
Claude Fable 5.2—Kimi K3.1 Pro90%SourceNot directly comparable
CharXiv
Claude Fable 5.2—Kimi K3.1 Pro93.4%SourceNot directly comparable
CharXiv w/o tools
Claude Fable 5.2—Kimi K3.1 Pro89%SourceNot directly comparable
BabyVision w/ Python
Claude Fable 5.2—Kimi K3.1 Pro95.4%SourceNot directly comparable
ZeroBench
Claude Fable 5.2—Kimi K3.1 Pro39.0%SourceNot directly comparable
ZeroBench w/ Python
Claude Fable 5.2—Kimi K3.1 Pro63.0%SourceNot directly comparable
GDP.pdf (Gemini judge)
Claude Fable 5.2—Kimi K3.1 Pro34.6%SourceNot directly comparable
GDP.pdf w/ Python (Gemini judge)
Claude Fable 5.2—Kimi K3.1 Pro85.7%SourceNot directly comparable
PerceptionBench
Claude Fable 5.2—Kimi K3.1 Pro68.3%SourceNot directly comparable
Knowledge
HLE
Claude Fable 5.274.2%SourceKimi K3.1 Pro70%SourceClaude Fable 5.2 leads this result
HLE w/o tools
Claude Fable 5.273.2%SourceKimi K3.1 Pro51.1%SourceClaude Fable 5.2 leads this result
PhysicianBench
Claude Fable 5.264.4%SourceKimi K3.1 Pro—Not directly comparable
GPQA
Claude Fable 5.2—Kimi K3.1 Pro94.2%SourceNot directly comparable
GPQA-D
Claude Fable 5.2—Kimi K3.1 Pro94.2%SourceNot directly comparable
Kimi Hallu Bench
Claude Fable 5.2—Kimi K3.1 Pro5.4%SourceNot directly comparable
Multilingual
69 public results · 8 shared
Watch Claude Fable 5.2 vs Kimi K3.1 Pro
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 6,100+ subscribers.
Last updated November 9, 2026