Kimi K2.7 Code 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.

Decision reading

Kimi K3 has the higher public point estimate, 72.14 versus 54.6. Their conditional score ranges overlap. These ranges do not establish rank confidence. 9 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.

Model A

Kimi K2.7 Code

Moonshot AI

54.6/100

Estimated · Public rank #71

Conditional range 40.3–69.0

Model B

Kimi K3

Moonshot AI

72.14/100

Supported · Public rank #15

90% interval 68.8–75.5

Shared results: 9
Kimi K2.7 Code only: 2
Kimi K3 only: 39
Like-for-like categories: 2 / 8

Estimated: Kimi K2.7 Code · 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.

Coding work
Code generation, repair, and software-engineering tasks
Kimi K3
Kimi K3 has the higher public coding point estimate, 61.4 to 44.5, with Supported evidence for both models. The conditional ranges do not establish rank confidence.
Confidence: limited
Agentic work
Tool use, computer use, and multi-step task completion
Kimi K3
Kimi K3 has the higher public agentic point estimate, 68.1 to 37.6, with Supported evidence for both models. The conditional ranges do not establish rank confidence.
Confidence: limited
Long documents
Prompts that approach the documented context limit
Kimi K3
Kimi K3 has the larger documented context window.
Confidence: documented

Show secondary and unsupported calls

Chat turn cost
1K fresh input + 500 output tokens
Kimi K2.7 Code
Kimi K2.7 Code has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-rates
Cache-heavy agent loop cost
200K cached + 20K fresh input + 10K output tokens
Kimi K2.7 Code
Kimi K2.7 Code has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-rates
Repository review cost
50K fresh input + 3K output tokens
Kimi K2.7 Code
Kimi K2.7 Code has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
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.

Pick the job you would use them for

The same task catalog as the LLM Selector. Each job names the evidence surface it rests on; nothing here adds a new score.

44.5Kimi K2.7 Code61.4Kimi K3

Like-for-like · BenchAlign v5.8

Kimi K3 has the higher coding point estimate. Conditional score ranges do not establish rank confidence.

Coding scores combine specific tasks and setups. Match the editor, harness, and effort to your workflow.

Open the coding leaderboard Shortlist models for this job

Same basis rules as the category table below

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.

CursorBench 3.2Coding
Normalized gap 11.1
Kimi K2.7 Code:49.7%
Kimi K3:60.8%
Shared source
LiveCodeBench (Vals)Coding
Normalized gap 5.1
Kimi K2.7 Code:82.1%
Kimi K3:87.2%
Kimi K2.7 Code source Kimi K3 source

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

Like-for-like

Kimi K2.7 Code: 37.6; Supported · #64/119
Kimi K3: 68.1; Supported · #8/119
Basis: BenchAlign v5.8 lane · 4 vs 12 public rows
Reading: Kimi K3 leads

Coding

Like-for-like

Kimi K2.7 Code: 44.5; Supported · #55/144
Kimi K3: 61.4; Supported · #18/144
Basis: BenchAlign v5.8 lane · 7 vs 14 public rows
Reading: Kimi K3 leads · intervals overlap

Knowledge

Directional only

Kimi K2.7 Code: 54.1; Estimated · #53/171
Kimi K3: 67.9; Supported · #18/171
Basis: BenchAlign v5.8 lane · 0 vs 6 public rows
Reading: Directional only

Reasoning

Not comparable

Kimi K2.7 Code: 76.8; Unranked · 2 rankable rows
Kimi K3: 65.8; #18/27
Basis: Provisional lane · 0 vs 1 weighted rows
Reading: Not comparable

Multimodal

Not comparable

Kimi K2.7 Code: Not ranked
Kimi K3: 89.4; #1/49
Basis: Provisional lane · 0 vs 3 weighted rows
Reading: Not comparable

Multilingual

Not comparable

Kimi K2.7 Code: Not ranked
Kimi K3: Not ranked
Basis: Provisional lane · 0 vs 0 weighted rows
Reading: Not comparable

Instruction following

Not comparable

Kimi K2.7 Code: 75.1; #59/125
Kimi K3: Not ranked
Basis: Provisional lane · 0 vs 0 weighted rows
Reading: Not comparable

Math

Not comparable

Kimi K2.7 Code: Not ranked
Kimi K3: Not ranked
Basis: Provisional lane · 0 vs 0 weighted rows
Reading: Not comparable

Category scores with the server-provided evidence basis for Kimi K2.7 Code and Kimi K3
Category	Kimi K2.7 Code	Kimi K3	Basis	Reading
Agentic	37.6Supported · #64/119	68.1Supported · #8/119	Like-for-likeBenchAlign v5.8 lane · 4 vs 12 public rows	Kimi K3 leads
Coding	44.5Supported · #55/144	61.4Supported · #18/144	Like-for-likeBenchAlign v5.8 lane · 7 vs 14 public rows	Kimi K3 leads · intervals overlap
Knowledge	54.1Estimated · #53/171	67.9Supported · #18/171	Directional onlyBenchAlign v5.8 lane · 0 vs 6 public rows	Directional only
Reasoning	76.8Unranked · 2 rankable rows	65.8#18/27	Not comparableProvisional lane · 0 vs 1 weighted rows	Not comparable
Multimodal	Not ranked	89.4#1/49	Not comparableProvisional lane · 0 vs 3 weighted rows	Not comparable
Multilingual	Not ranked	Not ranked	Not comparableProvisional lane · 0 vs 0 weighted rows	Not comparable
Instruction following	75.1#59/125	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, 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

Kimi K2.7 Code: $0.00295; Fits in one request
Kimi K3: $0.0105; Fits in one request

Kimi K2.7 Code has the lower modeled cost

Costs use the listed standard API rates.

Repository review

50K fresh input + 3K output tokens

Kimi K2.7 Code: $0.0595; Fits in one request
Kimi K3: $0.195; Fits in one request

Kimi K2.7 Code has the lower modeled cost

Costs use the listed standard API rates.

Cache-heavy agent loop

200K cached + 20K fresh input + 10K output tokens

Kimi K2.7 Code: $0.097; Fits in one request
Kimi K3: $0.27; Fits in one request

Kimi K2.7 Code 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.

SpecificationKimi K2.7 CodeKimi K3

Context window

Maximum documented context; output-token limits may be lower.

Kimi K2.7 Code

256K

Kimi K3

1.05M

API model ID

Kimi K2.7 Code

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.

Kimi K2.7 Code

$0.19 per 1M cached input tokens

Kimi K3

$0.3 per 1M cached input tokens

Documented inputs

Kimi K2.7 Code

Not sourced

Kimi K3

Not sourced

Documented outputs

Kimi K2.7 Code

Not sourced

Kimi K3

Not sourced

Provider availability

Kimi K2.7 Code

Not sourced

Kimi K3

Not sourced

Reasoning profile

Kimi K2.7 Code

Reasoning

Kimi K3

Reasoning

Weight access

Kimi K2.7 Code

Open Weight

Kimi K3

Pending

License

Kimi K2.7 Code

Open Weight

Kimi K3

Pending

Release date

Kimi K2.7 Code

2026-06-12

Kimi K3

2026-07-16

If you already use one of these models

Deployment change: Both entries list Moonshot AI as the provider. Confirm endpoint, model ID, limits, and feature support before switching.
Quality signal: Kimi K3 has the higher public point estimate, 72.14 versus 54.6. Their conditional score ranges overlap. These ranges do not establish rank confidence.
Workload cost: Repository review: $0.0595 vs $0.195. Cache-heavy agent loop: $0.097 vs $0.27.
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, Kimi K2.7 Code or Kimi K3?

Kimi K3 has the higher public point estimate, 72.14 versus 54.6. 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, Kimi K2.7 Code or Kimi K3?

Kimi K3 has the higher public coding point estimate, 61.4 to 44.5, with Supported evidence for both models. The conditional ranges do not establish rank confidence.

Which is better for agentic tasks, Kimi K2.7 Code or Kimi K3?

Kimi K3 has the higher public agentic tasks point estimate, 68.1 to 37.6, with Supported evidence for both models. The conditional ranges do not establish rank confidence.

Which costs less, Kimi K2.7 Code or Kimi K3?

For the stated presets, chat costs $0.00295 on Kimi K2.7 Code and $0.0105 on Kimi K3; repository review costs $0.0595 and $0.195; the cache-heavy agent loop costs $0.097 and $0.27. Costs use the listed standard API rates.

Which has the larger context window, Kimi K2.7 Code or Kimi K3?

Kimi K3 has the larger documented context window: 1.05M, compared with 256K.

Self-host vs API cost

Estimates at 50,000 req/day · 1000 tokens/req average.

Kimi K2.7 Code

API / mo$3,713

Self-host / mo$18,221

Break-even326M/day

Kimi K3

API / mo$0

Self-host / moNot listed

Break-even—

Proprietary model — self-hosting not applicable.

Model the full break-even

Benchmark evidence

The full public result ledger is available for audit without forcing a wide desktop table onto a phone.

Browse raw public benchmark evidence50 rows

Agentic

Kimi Claw 24/7
Kimi K2.7 Code46.9%
Source
Kimi K3—
Not directly comparable
MCP Atlas
Kimi K2.7 Code76%
Source
Kimi K384.2%
Source
Kimi K3 leads this result
MCP Mark Verified
Kimi K2.7 Code81.1%
Source
Kimi K3—
Not directly comparable
Terminal-Bench 2.1 (Vals)
Kimi K2.7 Code67.0%
Source
Kimi K380.9%
Source
Kimi K3 leads this result
Terminal-Bench 2.1
Kimi K2.7 Code—
Kimi K388.3%
Source
Not directly comparable
BrowseComp
Kimi K2.7 Code—
Kimi K391.2%
Source
Not directly comparable
DeepSearchQA
Kimi K2.7 Code—
Kimi K395.0%
Source
Not directly comparable
Toolathlon-Verified
Kimi K2.7 Code—
Kimi K373.2%
Source
Not directly comparable
AutomationBench
Kimi K2.7 Code—
Kimi K330.8%
Source
Not directly comparable
JobBench
Kimi K2.7 Code—
Kimi K352.9%
Source
Not directly comparable
APEX-Agents
Kimi K2.7 Code—
Kimi K337.6%
Source
Not directly comparable
SpreadsheetBench 2
Kimi K2.7 Code—
Kimi K334.8%
Source
Not directly comparable
DECK-Bench
Kimi K2.7 Code—
Kimi K373.5%
Source
Not directly comparable
ApprenticeBench
Kimi K2.7 Code—
Kimi K318%
Source
Not directly comparable

Coding

Kimi Code Bench v2
Kimi K2.7 Code62.0%
Source
Kimi K372.9%
Source
Kimi K3 leads this result
ProgramBench
Kimi K2.7 Code53.6%
Source
Kimi K377.8%
Source
Kimi K3 leads this result
MLS-Bench Lite
Kimi K2.7 Code35.1%
Source
Kimi K348.3%
Source
Kimi K3 leads this result
CursorBench 3.2
Shared source
Kimi K2.7 Code49.7%
Kimi K360.8%
Kimi K3 leads this result
OpenHarmony Bench
Shared source
Kimi K2.7 Code52.1%
Kimi K357.3%
Kimi K3 leads this result
LiveCodeBench (Vals)
Kimi K2.7 Code82.1%
Source
Kimi K387.2%
Source
Kimi K3 leads this result
SWE-bench (Vals)
Kimi K2.7 Code78.2%
Source
Kimi K393.4%
Source
Kimi K3 leads this result
DeepSWE
Kimi K2.7 Code—
Kimi K367.5%
Source
Not directly comparable
FrontierSWE
Kimi K2.7 Code—
Kimi K381.2%
Source
Not directly comparable
sweMarathon
Kimi K2.7 Code—
Kimi K342%
Source
Not directly comparable
PostTrain Bench
Kimi K2.7 Code—
Kimi K336.6%
Source
Not directly comparable
VulcanBench v3
Kimi K2.7 Code—
Kimi K373.7%
Source
Not directly comparable
FrontierSWE v2
Kimi K2.7 Code—
Kimi K325.9%
Source
Not directly comparable
PostTrainBench v1.1
Kimi K2.7 Code—
Kimi K332.0%
Source
Not directly comparable

Reasoning

ARC-AGI-1
Kimi K2.7 Code—
Kimi K394.50%
Source
Not directly comparable
ARC-AGI-2
Kimi K2.7 Code—
Kimi K360.4%
Source
Not directly comparable

Multimodal

OfficeQA Pro
Kimi K2.7 Code—
Kimi K363.3%
Source
Not directly comparable
MMMU-Pro
Kimi K2.7 Code—
Kimi K381.6%
Source
Not directly comparable
MMMU-Pro w/ Python
Kimi K2.7 Code—
Kimi K383.4%
Source
Not directly comparable
CharXiv w/o tools
Kimi K2.7 Code—
Kimi K384.8%
Source
Not directly comparable
CharXiv
Kimi K2.7 Code—
Kimi K391.3%
Source
Not directly comparable
MathVision
Kimi K2.7 Code—
Kimi K394.3%
Source
Not directly comparable
MathVision w/ Python
Kimi K2.7 Code—
Kimi K397.8%
Source
Not directly comparable
BabyVision w/ Python
Kimi K2.7 Code—
Kimi K385.7%
Source
Not directly comparable
ZeroBench
Kimi K2.7 Code—
Kimi K323.0%
Source
Not directly comparable
ZeroBench w/ Python
Kimi K2.7 Code—
Kimi K341.0%
Source
Not directly comparable
WorldVQA ForceAnswer
Kimi K2.7 Code—
Kimi K351.0%
Source
Not directly comparable
OmniDocBench
Kimi K2.7 Code—
Kimi K391.1%
Source
Not directly comparable
PerceptionBench
Kimi K2.7 Code—
Kimi K358.5%
Source
Not directly comparable

Knowledge

GPQA
Kimi K2.7 Code—
Kimi K393.5%
Source
Not directly comparable
GPQA-D
Kimi K2.7 Code—
Kimi K393.5%
Source
Not directly comparable
HLE
Kimi K2.7 Code—
Kimi K356%
Source
Not directly comparable
HLE w/o tools
Kimi K2.7 Code—
Kimi K343.5%
Source
Not directly comparable
GPQA Diamond (Vals)
Kimi K2.7 Code—
Kimi K392.9%
Source
Not directly comparable
MMLU-Pro (Vals)
Kimi K2.7 Code—
Kimi K388.0%
Source
Not directly comparable

Instruction following

Gray Swan IPI (15 attempts)
Kimi K2.7 Code—
Kimi K352.7%
Source
Not directly comparable

50 public results · 9 shared

Related comparisons

Compare API pricing Read the methodology Open the model selector

Watch Kimi K2.7 Code vs Kimi K3

One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.

Read a sample issue

Join 2,000+ readers.

Last updated October 2, 2026

Kimi K2.7 Code vs Kimi K3

Decision reading

Which one for your work

Coding work

Agentic work

Long documents

Chat turn cost

Cache-heavy agent loop cost

Repository review cost

Which one for a specific job

What is actually comparable

Shape of the matched evidence

Category results, on a stated basis

Agentic

Coding

Knowledge

Reasoning

Multimodal

Multilingual

Instruction following

Math

What each workload costs

Chat turn

Repository review

Cache-heavy agent loop

Specification differences

Context window

API model ID

Cached-input rate

Documented inputs

Documented outputs

Provider availability

Reasoning profile

Weight access

License

Release date

If you already use one of these models

Questions

Which is better, Kimi K2.7 Code or Kimi K3?

Which is better for coding, Kimi K2.7 Code or Kimi K3?

Which is better for agentic tasks, Kimi K2.7 Code or Kimi K3?

Which costs less, Kimi K2.7 Code or Kimi K3?

Which has the larger context window, Kimi K2.7 Code or Kimi K3?

Self-host vs API cost

Benchmark evidence

Agentic

Coding

Reasoning

Multimodal

Knowledge

Instruction following

Related comparisons

Watch Kimi K2.7 Code vs Kimi K3