Agentic
Directional only- Claude Sonnet 5
- 81.9
- Kimi K3
- 89.5
- Weighted basis
- 3 vs 2 rows
- Reading
- Directional only
Claude Sonnet 5 vs Kimi K3
Updated August 19, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.
Decision reading
Kimi K3 has the higher public score estimate, 80.53 versus 64.79, but the 90% score intervals overlap. Treat that as a lead, not a settled winner.
9 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
Claude Sonnet 5
Claude Sonnet 5 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
Claude Sonnet 5
Claude Sonnet 5 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
Claude Sonnet 5
Claude Sonnet 5 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
The category averages use different weighted benchmark sets, so they are directional rather than like-for-like.
Confidence: limited
What is actually comparable
Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.
10 Claude Sonnet 5 only
9 Shared
31 Kimi K3 only
- Shared results
- 9
- Claude Sonnet 5 only
- 10
- Kimi K3 only
- 31
- Like-for-like categories
- 0 / 8
3 categories use 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- Claude Sonnet 5
- 57.4
- Kimi K3
- 61.0
- Weighted basis
- 1 vs 2 rows
- Reading
- Directional only
Multimodal
Directional only- Claude Sonnet 5
- 88.3
- Kimi K3
- 78.5
- Weighted basis
- 1 vs 3 rows
- Reading
- Directional only
Coding
Not comparable- Claude Sonnet 5
- 76.7
- Kimi K3
- Not measured
- Weighted basis
- 2 vs 0 rows
- Reading
- Not comparable
Reasoning
Not comparable- Claude Sonnet 5
- Not measured
- Kimi K3
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
Math
Not comparable- Claude Sonnet 5
- Not measured
- Kimi K3
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
Multilingual
Not comparable- Claude Sonnet 5
- Not measured
- Kimi K3
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
Instruction following
Not comparable- Claude Sonnet 5
- Not measured
- Kimi K3
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
| Category | Claude Sonnet 5 | Kimi K3 | Weighted basis | Reading |
|---|---|---|---|---|
| Agentic | 81.9 | 89.5 | Directional only3 vs 2 rows | Directional only |
| Knowledge | 57.4 | 61.0 | Directional only1 vs 2 rows | Directional only |
| Multimodal | 88.3 | 78.5 | Directional only1 vs 3 rows | Directional only |
| Coding | 76.7 | Not measured | Not comparable2 vs 0 rows | Not comparable |
| Reasoning | Not measured | Not measured | Not comparable0 vs 0 rows | Not comparable |
| Math | Not measured | Not measured | Not comparable0 vs 0 rows | Not comparable |
| Multilingual | Not measured | Not measured | Not comparable0 vs 0 rows | Not comparable |
| Instruction following | Not measured | Not measured | Not comparable0 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.
Terminal-Bench 2.0
Agentic
BrowseComp
Agentic
CharXiv
Multimodal
HLE
Knowledge
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 Sonnet 5
- $0.007
- Fits in one request
- Kimi K3
- $0.0105
- Fits in one request
Claude Sonnet 5 has the lower modeled cost
Costs use the listed standard API rates.
Repository review
50K fresh input + 3K output tokens
- Claude Sonnet 5
- $0.13
- Fits in one request
- Kimi K3
- $0.195
- Fits in one request
Claude Sonnet 5 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 Sonnet 5
- $0.18
- Fits in one request
- Kimi K3
- $0.27
- Fits in one request
Claude Sonnet 5 has the lower modeled cost
Costs use the listed standard API rates.
Specification differences
Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.
SpecificationClaude Sonnet 5Kimi K3
Context window
Maximum documented context; output-token limits may be lower.
Claude Sonnet 5
Kimi K3
1.05M
API model ID
Claude Sonnet 5
claude-sonnet-5
Anthropic model overviewKimi K3
Not sourced
Cached-input rate
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
Claude Sonnet 5
$0.2 per 1M cached input tokens
Claude API pricingKimi K3
$0.3 per 1M cached input tokens
Documented inputs
Claude Sonnet 5
text, image
Anthropic model overviewKimi K3
Not sourced
Documented outputs
Claude Sonnet 5
Kimi K3
Not sourced
Provider availability
Claude Sonnet 5
Generally Available · Claude API
Anthropic model overviewKimi K3
Not sourced
Reasoning profile
Claude Sonnet 5
Reasoning
Kimi K3
Reasoning
Weight access
Claude Sonnet 5
Proprietary
Kimi K3
Pending
License
Claude Sonnet 5
Proprietary
Kimi K3
Pending
Release date
Claude Sonnet 5
2026-06-30
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 estimate, 80.53 versus 64.79, but the 90% score intervals overlap.
- Workload cost
- Repository review: $0.13 vs $0.195. Cache-heavy agent loop: $0.18 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.
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
Terminal-Bench 3.0
Claude Sonnet 514.6%SourceKimi K3—Not directly comparable
Terminal-Bench 2.0
Claude Sonnet 580.4%SourceKimi K388.3%SourceKimi K3 leads this result
BrowseComp
Claude Sonnet 584.7%SourceKimi K391.2%SourceKimi K3 leads this result
HLE w/ tools
Claude Sonnet 557.4%SourceKimi K3—Not directly comparable
OSWorld-Verified
Claude Sonnet 581.2%SourceKimi K3—Not directly comparable
DeepSearchQA
Claude Sonnet 5—Kimi K395.0%SourceNot directly comparable
Toolathlon-Verified
Claude Sonnet 5—Kimi K373.2%SourceNot directly comparable
MCP Atlas
Claude Sonnet 5—Kimi K384.2%SourceNot directly comparable
AutomationBench
Claude Sonnet 5—Kimi K330.8%SourceNot directly comparable
JobBench
Claude Sonnet 5—Kimi K352.9%SourceNot directly comparable
APEX-Agents
Claude Sonnet 5—Kimi K337.6%SourceNot directly comparable
SpreadsheetBench 2
Claude Sonnet 5—Kimi K334.8%SourceNot directly comparable
DECK-Bench
Claude Sonnet 5—Kimi K373.5%SourceNot directly comparable
Coding
SWE-bench Verified
Claude Sonnet 585.2%SourceKimi K3—Not directly comparable
SWE-bench Pro
Claude Sonnet 563.2%SourceKimi K3—Not directly comparable
SWE Multilingual
Claude Sonnet 578.3%SourceKimi K3—Not directly comparable
SWE Multimodal
Claude Sonnet 528.1%SourceKimi K3—Not directly comparable
Terminal-Bench 2.0
Claude Sonnet 580.4%SourceKimi K3—Not directly comparable
FrontierCode 1.1 Main
Claude Sonnet 542.7%SourceKimi K3—Not directly comparable
cursorBench32
Shared sourceClaude Sonnet 561.5%Kimi K360.8%Claude Sonnet 5 leads this result
APEX-SWE
Shared sourceClaude Sonnet 546.4%Kimi K348.0%Kimi K3 leads this result
EEBench
Shared sourceClaude Sonnet 540.3%Kimi K338.3%Claude Sonnet 5 leads this result
3DCodeBench
Claude Sonnet 539.2%SourceKimi K3—Not directly comparable
deepSwe
Claude Sonnet 5—Kimi K367.5%SourceNot directly comparable
FrontierSWE
Claude Sonnet 5—Kimi K381.2%SourceNot directly comparable
ProgramBench
Claude Sonnet 5—Kimi K377.8%SourceNot directly comparable
Kimi Code Bench v2
Claude Sonnet 5—Kimi K372.9%SourceNot directly comparable
sweMarathon
Claude Sonnet 5—Kimi K342%SourceNot directly comparable
PostTrain Bench
Claude Sonnet 5—Kimi K336.6%SourceNot directly comparable
MLS-Bench Lite
Claude Sonnet 5—Kimi K348.3%SourceNot directly comparable
VulcanBench v3
Claude Sonnet 5—Kimi K373.9%SourceNot directly comparable
InferenceEval
Claude Sonnet 5—Kimi K339.8%SourceNot directly comparable
KernelBench Internal
Claude Sonnet 5—Kimi K368.5%SourceNot directly comparable
Knowledge
Multimodal
CharXiv
Claude Sonnet 588.3%SourceKimi K391.3%SourceKimi K3 leads this result
CharXiv w/o tools
Claude Sonnet 577%SourceKimi K384.8%SourceKimi K3 leads this result
OfficeQA Pro
Claude Sonnet 5—Kimi K363.3%SourceNot directly comparable
MMMU-Pro
Claude Sonnet 5—Kimi K381.6%SourceNot directly comparable
MMMU-Pro w/ Python
Claude Sonnet 5—Kimi K383.4%SourceNot directly comparable
MathVision
Claude Sonnet 5—Kimi K394.3%SourceNot directly comparable
MathVision w/ Python
Claude Sonnet 5—Kimi K397.8%SourceNot directly comparable
BabyVision w/ Python
Claude Sonnet 5—Kimi K385.7%SourceNot directly comparable
ZeroBench
Claude Sonnet 5—Kimi K323.0%SourceNot directly comparable
ZeroBench w/ Python
Claude Sonnet 5—Kimi K341.0%SourceNot directly comparable
WorldVQA ForceAnswer
Claude Sonnet 5—Kimi K351.0%SourceNot directly comparable
OmniDocBench
Claude Sonnet 5—Kimi K391.1%SourceNot directly comparable
PerceptionBench
Claude Sonnet 5—Kimi K358.5%SourceNot directly comparable
Frequently asked questions
Which is better, Claude Sonnet 5 or Kimi K3?
Kimi K3 has the higher public score estimate, 80.53 versus 64.79, but the 90% score intervals overlap. The higher estimate is not a decisive winner because the uncertainty ranges overlap.
Which is better for coding, Claude Sonnet 5 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, Claude Sonnet 5 or Kimi K3?
The current agentic tasks averages use different weighted benchmark sets, so BenchLM does not name a winner from them. Read the shared benchmark rows directly and test the models on the same task set.
Which costs less, Claude Sonnet 5 or Kimi K3?
For the stated presets, chat costs $0.007 on Claude Sonnet 5 and $0.0105 on Kimi K3; repository review costs $0.13 and $0.195; the cache-heavy agent loop costs $0.18 and $0.27. Costs use the listed standard API rates.
Which has the larger context window, Claude Sonnet 5 or Kimi K3?
Kimi K3 has the larger documented context window: 1.05M, compared with 1M.
Related comparisons
Last updated August 19, 2026
Watch Claude Sonnet 5 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.