Model A

DeepSeek

44.3/100

Supported · Public rank #170

90% interval 25.4–63.2

DeepSeek V3 vs Kimi K3

Updated August 19, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.

Model B

Kimi K3

Moonshot AI

80.5/100

Supported · Public rank #5

90% interval 77.7–83.4

Decision reading

Kimi K3 has the higher public score, 80.53 versus 44.3, and the 90% score intervals do not overlap.

1 results are shared. Category rows based on different benchmark sets are marked directional and do not name a winner.

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.

Long documents
Prompts that approach the documented context limit
Kimi K3
Kimi K3 has the larger documented context window.
Confidence: documented
Chat turn cost
1K fresh input + 500 output tokens
DeepSeek V3
DeepSeek V3 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
DeepSeek V3
DeepSeek V3 has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-rates
Coding work
Code generation, repair, and software-engineering tasks
Not enough matched evidence
No shared weighted benchmark basis supports a winner.
Confidence: limited
Agentic work
Tool use, computer use, and multi-step task completion
Not enough matched evidence
No shared weighted benchmark basis supports a winner.
Confidence: limited
Cache-heavy agent loop cost
200K cached + 20K fresh input + 10K output tokens
Not enough matched evidence
The page does not recommend a cost winner because at least one model cannot fit the stated workload in one request. DeepSeek V3 does not fit this workload in one request.
Confidence: listed-rates

What is actually comparable

Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.

5 DeepSeek V3 only

1 Shared

39 Kimi K3 only

Shared results: 1
DeepSeek V3 only: 5
Kimi K3 only: 39
Like-for-like categories: 0 / 8

1 category uses different evidence sets. Those rows remain visible for coverage context but do not name a winner.

Category results, on a stated basis

Each row states whether both averages use the same weighted benchmark set. Directional and not-comparable rows remain visible, but they never receive a winner in this template.

Knowledge

Directional only

DeepSeek V3: 72.7
Kimi K3: 61.0
Weighted basis: 2 vs 2 rows
Reading: Directional only

Agentic

Not comparable

DeepSeek V3: Not measured
Kimi K3: 89.5
Weighted basis: 0 vs 2 rows
Reading: Not comparable

Coding

Not comparable

DeepSeek V3: 38.9
Kimi K3: Not measured
Weighted basis: 2 vs 0 rows
Reading: Not comparable

Reasoning

Not comparable

DeepSeek V3: Not measured
Kimi K3: Not measured
Weighted basis: 0 vs 0 rows
Reading: Not comparable

Math

Not comparable

DeepSeek V3: 1.7
Kimi K3: Not measured
Weighted basis: 1 vs 0 rows
Reading: Not comparable

Multilingual

Not comparable

DeepSeek V3: Not measured
Kimi K3: Not measured
Weighted basis: 0 vs 0 rows
Reading: Not comparable

Multimodal

Not comparable

DeepSeek V3: Not measured
Kimi K3: 78.5
Weighted basis: 0 vs 3 rows
Reading: Not comparable

Instruction following

Not comparable

DeepSeek V3: 86.1
Kimi K3: Not measured
Weighted basis: 1 vs 0 rows
Reading: Not comparable

Category averages with the server-provided evidence basis for DeepSeek V3 and Kimi K3
Category	DeepSeek V3	Kimi K3	Weighted basis	Reading
Knowledge	72.7	61.0	Directional only2 vs 2 rows	Directional only
Agentic	Not measured	89.5	Not comparable0 vs 2 rows	Not comparable
Coding	38.9	Not measured	Not comparable2 vs 0 rows	Not comparable
Reasoning	Not measured	Not measured	Not comparable0 vs 0 rows	Not comparable
Math	1.7	Not measured	Not comparable1 vs 0 rows	Not comparable
Multilingual	Not measured	Not measured	Not comparable0 vs 0 rows	Not comparable
Multimodal	Not measured	78.5	Not comparable0 vs 3 rows	Not comparable
Instruction following	86.1	Not measured	Not comparable1 vs 0 rows	Not comparable

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.

GPQA
Knowledge
DeepSeek V3: 59.1%Kimi K3: 93.5%Normalized gap 34.4DeepSeek V3 source Kimi K3 source

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

DeepSeek V3: $0.00082; Fits in one request
Kimi K3: $0.0105; Fits in one request

DeepSeek V3 has the lower modeled cost

Costs use the listed standard API rates.

Repository review

50K fresh input + 3K output tokens

DeepSeek V3: $0.0168; Fits in one request
Kimi K3: $0.195; Fits in one request

DeepSeek V3 has the lower modeled cost

Costs use the listed standard API rates.

Cache-heavy agent loop

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

DeepSeek V3: $0.0304; Does not fit in one request
Kimi K3: $0.27; Fits in one request

DeepSeek V3 does not fit this workload in one request.

Specification differences

Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.

SpecificationDeepSeek V3Kimi K3

Context window

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

DeepSeek V3

128K

Kimi K3

1.05M

API model ID

DeepSeek V3

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.

DeepSeek V3

$0.07 per 1M cached input tokens

Kimi K3

$0.3 per 1M cached input tokens

Documented inputs

DeepSeek V3

Not sourced

Kimi K3

Not sourced

Documented outputs

DeepSeek V3

Not sourced

Kimi K3

Not sourced

Provider availability

DeepSeek V3

Not sourced

Kimi K3

Not sourced

Reasoning profile

DeepSeek V3

Non-Reasoning

Kimi K3

Reasoning

Weight access

DeepSeek V3

Open Weight

Kimi K3

Pending

License

DeepSeek V3

Open Weight

Kimi K3

Pending

Release date

DeepSeek V3

2024-12-26

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 score, 80.53 versus 44.3, and the 90% score intervals do not overlap.
Workload cost: Repository review: $0.0168 vs $0.195. Cache-heavy agent loop: $0.0304 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.

Self-host vs API cost

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

DeepSeek V3

API / mo$1,028

Self-host / mo$18,221

Break-even1.2B/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 evidence45 rows

Agentic

Terminal-Bench 2.0
DeepSeek V3—
Kimi K388.3%
Source
Not directly comparable
BrowseComp
DeepSeek V3—
Kimi K391.2%
Source
Not directly comparable
DeepSearchQA
DeepSeek V3—
Kimi K395.0%
Source
Not directly comparable
Toolathlon-Verified
DeepSeek V3—
Kimi K373.2%
Source
Not directly comparable
MCP Atlas
DeepSeek V3—
Kimi K384.2%
Source
Not directly comparable
AutomationBench
DeepSeek V3—
Kimi K330.8%
Source
Not directly comparable
JobBench
DeepSeek V3—
Kimi K352.9%
Source
Not directly comparable
APEX-Agents
DeepSeek V3—
Kimi K337.6%
Source
Not directly comparable
SpreadsheetBench 2
DeepSeek V3—
Kimi K334.8%
Source
Not directly comparable
DECK-Bench
DeepSeek V3—
Kimi K373.5%
Source
Not directly comparable

Coding

LiveCodeBench
DeepSeek V337.6%
Source
Kimi K3—
Not directly comparable
SWE-bench Verified
DeepSeek V342%
Source
Kimi K3—
Not directly comparable
deepSwe
DeepSeek V3—
Kimi K367.5%
Source
Not directly comparable
cursorBench32
DeepSeek V3—
Kimi K360.8%
Source
Not directly comparable
FrontierSWE
DeepSeek V3—
Kimi K381.2%
Source
Not directly comparable
ProgramBench
DeepSeek V3—
Kimi K377.8%
Source
Not directly comparable
Kimi Code Bench v2
DeepSeek V3—
Kimi K372.9%
Source
Not directly comparable
sweMarathon
DeepSeek V3—
Kimi K342%
Source
Not directly comparable
PostTrain Bench
DeepSeek V3—
Kimi K336.6%
Source
Not directly comparable
MLS-Bench Lite
DeepSeek V3—
Kimi K348.3%
Source
Not directly comparable
VulcanBench v3
DeepSeek V3—
Kimi K373.9%
Source
Not directly comparable
APEX-SWE
DeepSeek V3—
Kimi K348.0%
Source
Not directly comparable
EEBench
DeepSeek V3—
Kimi K338.3%
Source
Not directly comparable
InferenceEval
DeepSeek V3—
Kimi K339.8%
Source
Not directly comparable
KernelBench Internal
DeepSeek V3—
Kimi K368.5%
Source
Not directly comparable

Knowledge

GPQA
DeepSeek V359.1%
Source
Kimi K393.5%
Source
Kimi K3 leads this result
MMLU-Pro
DeepSeek V375.9%
Source
Kimi K3—
Not directly comparable
GPQA-D
DeepSeek V3—
Kimi K393.5%
Source
Not directly comparable
HLE
DeepSeek V3—
Kimi K356%
Source
Not directly comparable
HLE w/o tools
DeepSeek V3—
Kimi K343.5%
Source
Not directly comparable

Math

FrontierMath v2 (Tiers 1-3)
DeepSeek V31.724%
Source
Kimi K3—
Not directly comparable

Multimodal

OfficeQA Pro
DeepSeek V3—
Kimi K363.3%
Source
Not directly comparable
MMMU-Pro
DeepSeek V3—
Kimi K381.6%
Source
Not directly comparable
MMMU-Pro w/ Python
DeepSeek V3—
Kimi K383.4%
Source
Not directly comparable
CharXiv w/o tools
DeepSeek V3—
Kimi K384.8%
Source
Not directly comparable
CharXiv
DeepSeek V3—
Kimi K391.3%
Source
Not directly comparable
MathVision
DeepSeek V3—
Kimi K394.3%
Source
Not directly comparable
MathVision w/ Python
DeepSeek V3—
Kimi K397.8%
Source
Not directly comparable
BabyVision w/ Python
DeepSeek V3—
Kimi K385.7%
Source
Not directly comparable
ZeroBench
DeepSeek V3—
Kimi K323.0%
Source
Not directly comparable
ZeroBench w/ Python
DeepSeek V3—
Kimi K341.0%
Source
Not directly comparable
WorldVQA ForceAnswer
DeepSeek V3—
Kimi K351.0%
Source
Not directly comparable
OmniDocBench
DeepSeek V3—
Kimi K391.1%
Source
Not directly comparable
PerceptionBench
DeepSeek V3—
Kimi K358.5%
Source
Not directly comparable

Instruction following

IFEval
DeepSeek V386.1%
Source
Kimi K3—
Not directly comparable

Frequently asked questions

Which is better, DeepSeek V3 or Kimi K3?

Kimi K3 has the higher public score, 80.53 versus 44.3, and the 90% score intervals do not overlap. That is the clearest overall quality signal in the current public data.

Which is better for coding, DeepSeek V3 or Kimi K3?

The published evidence does not provide a shared weighted coding basis for both models, so BenchLM does not name a coding winner.

Which is better for agentic tasks, DeepSeek V3 or Kimi K3?

The published evidence does not provide a shared weighted agentic tasks basis for both models, so BenchLM does not name a agentic tasks winner.

Which costs less, DeepSeek V3 or Kimi K3?

For the stated presets, chat costs $0.00082 on DeepSeek V3 and $0.0105 on Kimi K3; repository review costs $0.0168 and $0.195; the cache-heavy agent loop costs $0.0304 and $0.27. DeepSeek V3 does not fit this workload in one request.

Which has the larger context window, DeepSeek V3 or Kimi K3?

Kimi K3 has the larger documented context window: 1.05M, compared with 128K.

Related comparisons

Compare API pricing Read the methodology Open the model selector

Last updated August 19, 2026

Watch DeepSeek V3 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.

DeepSeek V3 vs Kimi K3

Which one for your work

Long documents

Chat turn cost

Repository review cost

Coding work

Agentic work

Cache-heavy agent loop cost

What is actually comparable

Category results, on a stated basis

Knowledge

Agentic

Coding

Reasoning

Math

Multilingual

Multimodal

Instruction following

Shape of the matched evidence

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

Self-host vs API cost

Benchmark evidence

Agentic

Coding

Knowledge

Math

Multimodal

Instruction following

Frequently asked questions

Which is better, DeepSeek V3 or Kimi K3?

Which is better for coding, DeepSeek V3 or Kimi K3?

Which is better for agentic tasks, DeepSeek V3 or Kimi K3?

Which costs less, DeepSeek V3 or Kimi K3?

Which has the larger context window, DeepSeek V3 or Kimi K3?

Related comparisons

Watch DeepSeek V3 vs Kimi K3