Agentic work
Tool use, computer use, and multi-step task completion
Kimi K3
Kimi K3 has the higher public agentic point estimate, 65.1 to 36, with Supported evidence for both models. The conditional ranges do not establish rank confidence.
Updated November 17, 2026. Rank says Kimi K3 is ahead. Price, access, and your workload can each overturn that. Public scores include evidence status and uncertainty.
Kimi K3 has the higher public point estimate, 67 versus 46.36. Their conditional score ranges do not overlap. These ranges do not establish rank confidence. 3 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.
Claude Sonnet 4.5: $3 input / $15 output
Kimi K3: $3 input / $15 output
Claude Sonnet 4.5: 200K
Kimi K3: 1.05M
Claude Sonnet 4.5: September 29, 2025
Kimi K3: July 16, 2026
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.
Tool use, computer use, and multi-step task completion
Kimi K3
Kimi K3 has the higher public agentic point estimate, 65.1 to 36, with Supported evidence for both models. The conditional ranges do not establish rank confidence.
Prompts that approach the documented context limit
Kimi K3
Kimi K3 has the larger documented context window.
Code generation, repair, and software-engineering tasks
Not enough matched evidence
Claude Sonnet 4.5 is not ranked on the public lane for coding, so no winner is named for coding.
1K fresh input + 500 output tokens
No clear pick
A complete comparable API-rate estimate is not available for both models.
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. Claude Sonnet 4.5 does not fit this workload in one request. Claude Sonnet 4.5 has no published cached-input rate, so cached tokens use its listed input rate.
50K fresh input + 3K output tokens
No clear pick
A complete comparable API-rate estimate is not available for both models.
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.
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.
The LLM Selector reads what you need a model for and returns a shortlist, ranked in the same order the job pages publish.
We keep what you type here, drafts included, for 180 days, and after that only the job and requirements read from it. Nothing is tied to your IP address or account. Leave out names and other personal details. Privacy
Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.
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 46.8GPQAKnowledge
Normalized gap 10.1Each 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.
| Category | Claude Sonnet 4.5 | Kimi K3 | Basis | Reading |
|---|---|---|---|---|
| Agentic | 36.0Supported · #86/140 | 65.1Supported · #21/140 | Like-for-likeBenchAlign v5.8 lane · 5 vs 19 public rows | Kimi K3 leads |
| Coding | Not ranked | 56.3Supported · #23/163 | Not comparableBenchAlign v5.8 lane · 2 vs 18 public rows | Not comparable |
| Reasoning | Not ranked | 65.4Supported · #21/161 | Not comparableBenchAlign v5.8 lane · 1 vs 2 public rows | Not comparable |
| Multimodal | Not ranked | 66.7Estimated · #19/67 | Not comparableBenchAlign v5.8 lane · 0 vs 14 public rows | Not comparable |
| Knowledge | Not ranked | 67.5Supported · #20/181 | 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 | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Math | 32.4Unranked · 2 rankable rows | Not ranked | Not comparableProvisional lane · 2 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.
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.
1K fresh input + 500 output tokens
Modeled costs are equal
Costs use the listed standard API rates.
50K fresh input + 3K output tokens
Modeled costs are equal
Costs use the listed standard API rates.
200K cached + 20K fresh input + 10K output tokens
Claude Sonnet 4.5 does not fit this workload in one request. Claude Sonnet 4.5 has no published cached-input rate, so cached tokens use its listed input rate.
Enter a month of tokens to price it at both models’ listed API rates.
Standard listed rates, no batch or volume discounts.
Overall score against output price per 1M tokens, one model per family. Pricing as of November 17, 2026.
Claude Sonnet 4.5Kimi K3
Gray dots are the other priced models. The dark step line is the efficiency frontier: for a model under it, another model costs no more and scores at least as high.
See Claude Sonnet 4.5 on the full price vs performance chartSourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.
Maximum documented context; output-token limits may be lower.
Claude Sonnet 4.5
200K
Kimi K3
1.05M
Claude Sonnet 4.5
Not sourced
Kimi K3
Not sourced
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
Claude Sonnet 4.5
Not published
Kimi K3
$0.3 per 1M cached input tokens
Claude Sonnet 4.5
text, image
Anthropic Claude Sonnet 4.5 model pageKimi K3
text, image, video
Kimi-K3 model cardClaude Sonnet 4.5
Kimi K3
Claude Sonnet 4.5
Not sourced
Kimi K3
Not sourced
Claude Sonnet 4.5
Non-Reasoning
Kimi K3
Reasoning
Claude Sonnet 4.5
Proprietary
Kimi K3
Open Weight
Claude Sonnet 4.5
Proprietary
Kimi K3
Open Weight
Claude Sonnet 4.5
2025-09-29
Kimi K3
2026-07-16
Kimi K3 has the higher public point estimate, 67 versus 46.36. Their conditional score ranges do not overlap. These ranges do not establish rank confidence. The page therefore keeps the decision tied to the specific documented workload.
Claude Sonnet 4.5 is not ranked on the public lane for coding, so no winner is named for coding.
Kimi K3 has the higher public agentic tasks point estimate, 65.1 to 36, with Supported evidence for both models. The conditional ranges do not establish rank confidence.
For the stated presets, chat costs $0.0105 on Claude Sonnet 4.5 and $0.0105 on Kimi K3; repository review costs $0.195 and $0.195; the cache-heavy agent loop costs $0.81 and $0.27. Claude Sonnet 4.5 does not fit this workload in one request. Claude Sonnet 4.5 has no published cached-input rate, so cached tokens use its listed input rate.
Kimi K3 has the larger documented context window: 1.05M, compared with 200K.
Estimates at 50,000 req/day · 1000 tokens/req average.
The full public result ledger is available for audit without forcing a wide desktop table onto a phone.
Terminal-Bench 2.0
Not directly comparable
OSWorld-Verified
Not directly comparable
VITA-Bench
Not directly comparable
Gert Labs
Not directly comparable
JobBench
Kimi K3 leads this result
Terminal-Bench 4.1
Not directly comparable
Finch
Not directly comparable
b3 Agent Security
Not directly comparable
Cyber Misuse Refusal
Not directly comparable
Terminal-Bench 2.1
Not directly comparable
BrowseComp
Not directly comparable
DeepSearchQA
Not directly comparable
Toolathlon-Verified
Not directly comparable
MCP Atlas
Not directly comparable
AutomationBench
Not directly comparable
APEX-Agents
Not directly comparable
SpreadsheetBench 2
Not directly comparable
DECK-Bench
Not directly comparable
Terminal-Bench 2.1 (Vals)
Not directly comparable
ApprenticeBench
Not directly comparable
Hermes Index
Not directly comparable
Hermes Bench
Not directly comparable
MazeBench
Not directly comparable
SWE-bench Verified
Not directly comparable
Terminal-Bench 2.0
Not directly comparable
ScarfBench
Not directly comparable
DeepSWE
Not directly comparable
CursorBench 3.2
Not directly comparable
FrontierSWE
Not directly comparable
ProgramBench
Not directly comparable
Kimi Code Bench v2
Not directly comparable
sweMarathon
Not directly comparable
PostTrainBench (lab runs)
Not directly comparable
MLS-Bench Lite
Not directly comparable
VulcanBench v3
Not directly comparable
OpenHarmony Bench v1.0
Not directly comparable
FrontierSWE v2
Not directly comparable
LiveCodeBench (Vals)
Not directly comparable
SWE-bench (Vals)
Not directly comparable
PostTrainBench v1.1
Not directly comparable
Bug Hunt Bench
Not directly comparable
FrontierCode 2.0
Not directly comparable
PostTrainBench
Not directly comparable
ExtractBench (LlamaIndex)
Not directly comparable
OfficeQA Pro
Not directly comparable
MMMU-Pro
Not directly comparable
MMMU-Pro w/ Python
Not directly comparable
CharXiv w/o tools
Not directly comparable
CharXiv
Not directly comparable
MathVision
Not directly comparable
MathVision w/ Python
Not directly comparable
BabyVision w/ Python
Not directly comparable
ZeroBench
Not directly comparable
ZeroBench w/ Python
Not directly comparable
WorldVQA ForceAnswer
Not directly comparable
OmniDocBench
Not directly comparable
PerceptionBench
Not directly comparable
GPQA
Kimi K3 leads this result
GPQA-D
Not directly comparable
HLE
Not directly comparable
HLE w/o tools
Not directly comparable
GPQA Diamond (Vals)
Not directly comparable
MMLU-Pro (Vals)
Not directly comparable
Gray Swan IPI (15 attempts)
Not directly comparable
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 7,000+ subscribers.
Last updated November 17, 2026