Long documents
Prompts that approach the documented context limit
Kimi K3
Kimi K3 has the larger documented context window.
Confidence: documented
dots3-note Preview vs Kimi K3
Updated October 2, 2026. Rank says Kimi K3 is ahead. Price, access, and your workload can each overturn that. Public scores include evidence status and uncertainty.
Both of these models will change. Get the price, version and retirement notices for the pair, each with its source. Follow model changes
Share or export
Decision reading
Kimi K3 has the higher public point estimate, 72.14 versus 63.63. Their conditional score ranges overlap. These ranges do not establish rank confidence. 12 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.
- Shared results
- 12
- dots3-note Preview only
- 19
- Kimi K3 only
- 36
- Like-for-like categories
- 0 / 8
Estimated: dots3-note Preview · Supported: Kimi K3. 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.
Show secondary and unsupported calls
Coding work
Code generation, repair, and software-engineering tasks
Not enough matched evidence
Dots3-note Preview 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
Dots3-note Preview is scored on Estimated evidence for agentic, so the reading is directional rather than like-for-like.
Confidence: limitedChat turn cost
1K fresh input + 500 output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
Confidence: listed-ratesCache-heavy agent loop cost
200K cached + 20K fresh input + 10K output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
Confidence: listed-ratesRepository review cost
50K fresh input + 3K output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
Confidence: listed-rates
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.
—dots3-note Preview61.4Kimi K3
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.
19 dots3-note Preview only
12 Shared
36 Kimi K3 only
1 category rests 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.
ARC-AGI-2Reasoning
Normalized gap 21.0dots3-note Preview:81.4%Kimi K3:60.4%BrowseCompAgentic
Normalized gap 7.9dots3-note Preview:83.3%Kimi K3:91.2%HLEKnowledge
Normalized gap 3.4dots3-note Preview:52.6%Kimi K3:56%MMMU-ProMultimodal
Normalized gap 2.5dots3-note Preview:79.1%Kimi K3:81.6%
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, 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.
Agentic
Directional only- dots3-note Preview
- 56.7
- Estimated · #30/119
- Kimi K3
- 68.1
- Supported · #8/119
- Basis
- BenchAlign v5.8 lane · 9 vs 12 public rows
- Reading
- Directional only
Coding
Not comparable- dots3-note Preview
- Not ranked
- Kimi K3
- 61.4
- Supported · #18/144
- Basis
- BenchAlign v5.8 lane · 7 vs 14 public rows
- Reading
- Not comparable
Reasoning
Not comparable- dots3-note Preview
- 72.8
- Unranked · 1 rankable row
- Kimi K3
- 65.8
- #18/27
- Basis
- Provisional lane · 1 vs 1 weighted rows
- Reading
- Not comparable
Multimodal
Not comparable- dots3-note Preview
- 67.2
- Unranked · 10 rankable rows
- Kimi K3
- 89.4
- #1/49
- Basis
- Provisional lane · 1 vs 3 weighted rows
- Reading
- Not comparable
Knowledge
Not comparable- dots3-note Preview
- Not ranked
- Kimi K3
- 67.9
- Supported · #18/171
- Basis
- BenchAlign v5.8 lane · 1 vs 6 public rows
- Reading
- Not comparable
Multilingual
Not comparable- dots3-note Preview
- Not ranked
- Kimi K3
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Instruction following
Not comparable- dots3-note Preview
- 84.5
- #39/125
- Kimi K3
- Not ranked
- Basis
- Provisional lane · 1 vs 0 weighted rows
- Reading
- Not comparable
Math
Not comparable- dots3-note Preview
- Not ranked
- Kimi K3
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
| Category | dots3-note Preview | Kimi K3 | Basis | Reading |
|---|---|---|---|---|
| Agentic | 56.7Estimated · #30/119 | 68.1Supported · #8/119 | Directional onlyBenchAlign v5.8 lane · 9 vs 12 public rows | Directional only |
| Coding | Not ranked | 61.4Supported · #18/144 | Not comparableBenchAlign v5.8 lane · 7 vs 14 public rows | Not comparable |
| Reasoning | 72.8Unranked · 1 rankable row | 65.8#18/27 | Not comparableProvisional lane · 1 vs 1 weighted rows | Not comparable |
| Multimodal | 67.2Unranked · 10 rankable rows | 89.4#1/49 | Not comparableProvisional lane · 1 vs 3 weighted rows | Not comparable |
| Knowledge | Not ranked | 67.9Supported · #18/171 | Not comparableBenchAlign v5.8 lane · 1 vs 6 public rows | Not comparable |
| Multilingual | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Instruction following | 84.5#39/125 | Not ranked | Not comparableProvisional lane · 1 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, 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
- dots3-note Preview
- Self-hosted; infrastructure cost varies
- Fits in one request
- Kimi K3
- $0.0105
- Fits in one request
dots3-note Preview has no comparable published API token rate.
Repository review
50K fresh input + 3K output tokens
- dots3-note Preview
- Self-hosted; infrastructure cost varies
- Fits in one request
- Kimi K3
- $0.195
- Fits in one request
dots3-note Preview has no comparable published API token rate.
Cache-heavy agent loop
200K cached + 20K fresh input + 10K output tokens
- dots3-note Preview
- Self-hosted; infrastructure cost varies
- Fits in one request
- Cached-input rate unavailable
- Kimi K3
- $0.27
- Fits in one request
dots3-note Preview has no comparable published API token rate.
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.
Specificationdots3-note PreviewKimi K3
Context window
Maximum documented context; output-token limits may be lower.
dots3-note Preview
Kimi K3
1.05M
API model ID
dots3-note Preview
Not sourced
Kimi K3
Not sourced
Cached-input rate
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
dots3-note Preview
No comparable hosted API rate
dots3-note Preview model cardKimi K3
$0.3 per 1M cached input tokens
Documented inputs
dots3-note Preview
Not sourced
Kimi K3
Not sourced
Documented outputs
dots3-note Preview
Not sourced
Kimi K3
Not sourced
Provider availability
dots3-note Preview
Not sourced
Kimi K3
Not sourced
Reasoning profile
dots3-note Preview
Reasoning
Kimi K3
Reasoning
Weight access
dots3-note Preview
Open Weight
Kimi K3
Pending
License
dots3-note Preview
Open Weight
Kimi K3
Pending
Release date
dots3-note Preview
2026-08-14
Kimi K3
2026-07-16
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
- Kimi K3 has the higher public point estimate, 72.14 versus 63.63. Their conditional score ranges overlap. These ranges do not establish rank confidence.
- Workload cost
- A complete comparable API-rate estimate is not available for both models.
- Context tradeoff
- Kimi K3 has the larger documented window (1.05M).
Run the same representative tasks against both endpoints before changing production traffic.
Questions
Which is better, dots3-note Preview or Kimi K3?
Kimi K3 has the higher public point estimate, 72.14 versus 63.63. 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, dots3-note Preview or Kimi K3?
Dots3-note Preview is not ranked on the public lane for coding, so no winner is named for coding.
Which is better for agentic tasks, dots3-note Preview or Kimi K3?
Kimi K3 scores higher for agentic tasks on the public lane, 68.1 to 56.7. Dots3-note Preview 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, dots3-note Preview or Kimi K3?
Both models do not have comparable published API token rates, so this page does not name a universal price winner.
Which has the larger context window, dots3-note Preview or Kimi K3?
Kimi K3 has the larger documented context window: 1.05M, compared with 512K.
Benchmark evidence
The full public result ledger is available for audit without forcing a wide desktop table onto a phone.
Browse raw public benchmark evidence67 rows
Agentic
Claw-Eval
dots3-note Preview73.4%SourceKimi K3—Not directly comparable
Terminal-Bench 2.1
dots3-note Preview75.1%SourceKimi K388.3%SourceKimi K3 leads this result
Toolathlon-Verified
dots3-note Preview55.6%SourceKimi K373.2%SourceKimi K3 leads this result
skillsBench
dots3-note Preview52.8%SourceKimi K3—Not directly comparable
APEX-Agents
dots3-note Preview30.8%SourceKimi K337.6%SourceKimi K3 leads this result
BrowseComp
dots3-note Preview83.3%SourceKimi K391.2%SourceKimi K3 leads this result
HLE w/ tools
dots3-note Preview52.6%SourceKimi K3—Not directly comparable
DeepSearchQA
dots3-note Preview92.1%SourceKimi K395.0%SourceKimi K3 leads this result
WideResearch
dots3-note Preview78.9%SourceKimi K3—Not directly comparable
MCP Atlas
dots3-note Preview—Kimi K384.2%SourceNot directly comparable
AutomationBench
dots3-note Preview—Kimi K330.8%SourceNot directly comparable
JobBench
dots3-note Preview—Kimi K352.9%SourceNot directly comparable
SpreadsheetBench 2
dots3-note Preview—Kimi K334.8%SourceNot directly comparable
DECK-Bench
dots3-note Preview—Kimi K373.5%SourceNot directly comparable
Terminal-Bench 2.1 (Vals)
dots3-note Preview—Kimi K380.9%SourceNot directly comparable
ApprenticeBench
dots3-note Preview—Kimi K318%SourceNot directly comparable
Coding
Codeforces
dots3-note Preview3056.0SourceKimi K3—Not directly comparable
LiveCodeBench v6
dots3-note Preview91.5%SourceKimi K3—Not directly comparable
Terminal-Bench 2.1
dots3-note Preview75.1%SourceKimi K3—Not directly comparable
SWE-bench Verified
dots3-note Preview78.4%SourceKimi K3—Not directly comparable
SWE Multilingual
dots3-note Preview75.7%SourceKimi K3—Not directly comparable
SWE-bench Pro
dots3-note Preview61%SourceKimi K3—Not directly comparable
NL2Repo
dots3-note Preview49.8%SourceKimi K3—Not directly comparable
DeepSWE
dots3-note Preview—Kimi K367.5%SourceNot directly comparable
CursorBench 3.2
dots3-note Preview—Kimi K360.8%SourceNot directly comparable
FrontierSWE
dots3-note Preview—Kimi K381.2%SourceNot directly comparable
ProgramBench
dots3-note Preview—Kimi K377.8%SourceNot directly comparable
Kimi Code Bench v2
dots3-note Preview—Kimi K372.9%SourceNot directly comparable
sweMarathon
dots3-note Preview—Kimi K342%SourceNot directly comparable
PostTrain Bench
dots3-note Preview—Kimi K336.6%SourceNot directly comparable
MLS-Bench Lite
dots3-note Preview—Kimi K348.3%SourceNot directly comparable
VulcanBench v3
dots3-note Preview—Kimi K373.7%SourceNot directly comparable
OpenHarmony Bench
dots3-note Preview—Kimi K357.3%SourceNot directly comparable
FrontierSWE v2
dots3-note Preview—Kimi K325.9%SourceNot directly comparable
LiveCodeBench (Vals)
dots3-note Preview—Kimi K387.2%SourceNot directly comparable
SWE-bench (Vals)
dots3-note Preview—Kimi K393.4%SourceNot directly comparable
PostTrainBench v1.1
dots3-note Preview—Kimi K332.0%SourceNot directly comparable
Reasoning
Multimodal
SimpleVQA
dots3-note Preview72.5%SourceKimi K3—Not directly comparable
MMMU-Pro
dots3-note Preview79.1%SourceKimi K381.6%SourceKimi K3 leads this result
MathVision
dots3-note Preview87.7%SourceKimi K394.3%SourceKimi K3 leads this result
ZeroBench
dots3-note Preview19.0%SourceKimi K323.0%SourceKimi K3 leads this result
CharXiv w/o tools
dots3-note Preview83.1%SourceKimi K384.8%SourceKimi K3 leads this result
GDP.pdf (no tools)
dots3-note Preview60.7%SourceKimi K3—Not directly comparable
PerceptionBench
dots3-note Preview53.4%SourceKimi K358.5%SourceKimi K3 leads this result
BabyVision
dots3-note Preview50.0%SourceKimi K3—Not directly comparable
MMVU
dots3-note Preview79.9%SourceKimi K3—Not directly comparable
VideoMMMU
dots3-note Preview86.8%SourceKimi K3—Not directly comparable
OfficeQA Pro
dots3-note Preview—Kimi K363.3%SourceNot directly comparable
MMMU-Pro w/ Python
dots3-note Preview—Kimi K383.4%SourceNot directly comparable
CharXiv
dots3-note Preview—Kimi K391.3%SourceNot directly comparable
MathVision w/ Python
dots3-note Preview—Kimi K397.8%SourceNot directly comparable
BabyVision w/ Python
dots3-note Preview—Kimi K385.7%SourceNot directly comparable
ZeroBench w/ Python
dots3-note Preview—Kimi K341.0%SourceNot directly comparable
WorldVQA ForceAnswer
dots3-note Preview—Kimi K351.0%SourceNot directly comparable
OmniDocBench
dots3-note Preview—Kimi K391.1%SourceNot directly comparable
Knowledge
HLE
dots3-note Preview52.6%SourceKimi K356%SourceKimi K3 leads this result
GPQA
dots3-note Preview—Kimi K393.5%SourceNot directly comparable
GPQA-D
dots3-note Preview—Kimi K393.5%SourceNot directly comparable
HLE w/o tools
dots3-note Preview—Kimi K343.5%SourceNot directly comparable
GPQA Diamond (Vals)
dots3-note Preview—Kimi K392.9%SourceNot directly comparable
MMLU-Pro (Vals)
dots3-note Preview—Kimi K388.0%SourceNot directly comparable
Instruction following
Math
IMOAnswerBench
dots3-note Preview90.9%SourceKimi K3—Not directly comparable
67 public results · 12 shared
Watch dots3-note Preview vs Kimi K3
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 2,000+ readers.
Last updated October 2, 2026