Knowledge
Like-for-like- GPT-5.4 mini
- 47.8
- Kimi K3
- 61.0
- Weighted basis
- 2 vs 2 rows
- Reading
- Kimi K3 leads
GPT-5.4 mini 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, 80.53 versus 56.85, and the 90% score intervals do not overlap.
8 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
GPT-5.4 mini
GPT-5.4 mini 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
GPT-5.4 mini
GPT-5.4 mini 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
GPT-5.4 mini
GPT-5.4 mini 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.
7 GPT-5.4 mini only
8 Shared
32 Kimi K3 only
- Shared results
- 8
- GPT-5.4 mini only
- 7
- Kimi K3 only
- 32
- Like-for-like categories
- 1 / 8
2 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.
Agentic
Directional only- GPT-5.4 mini
- 65.7
- Kimi K3
- 89.5
- Weighted basis
- 2 vs 2 rows
- Reading
- Directional only
Multimodal
Directional only- GPT-5.4 mini
- 76.6
- Kimi K3
- 78.5
- Weighted basis
- 1 vs 3 rows
- Reading
- Directional only
Coding
Not comparable- GPT-5.4 mini
- Not measured
- Kimi K3
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
Reasoning
Not comparable- GPT-5.4 mini
- Not measured
- Kimi K3
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
Math
Not comparable- GPT-5.4 mini
- 21.7
- Kimi K3
- Not measured
- Weighted basis
- 2 vs 0 rows
- Reading
- Not comparable
Multilingual
Not comparable- GPT-5.4 mini
- Not measured
- Kimi K3
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
Instruction following
Not comparable- GPT-5.4 mini
- Not measured
- Kimi K3
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
| Category | GPT-5.4 mini | Kimi K3 | Weighted basis | Reading |
|---|---|---|---|---|
| Knowledge | 47.8 | 61.0 | Like-for-like2 vs 2 rows | Kimi K3 leads |
| Agentic | 65.7 | 89.5 | Directional only2 vs 2 rows | Directional only |
| Multimodal | 76.6 | 78.5 | Directional only1 vs 3 rows | Directional only |
| Coding | Not measured | Not measured | Not comparable0 vs 0 rows | Not comparable |
| Reasoning | Not measured | Not measured | Not comparable0 vs 0 rows | Not comparable |
| Math | 21.7 | Not measured | Not comparable2 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
HLE
Knowledge
GPQA
Knowledge
MMMU-Pro
Multimodal
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
- GPT-5.4 mini
- $0.003
- Fits in one request
- Kimi K3
- $0.0105
- Fits in one request
GPT-5.4 mini has the lower modeled cost
Costs use the listed standard API rates.
Repository review
50K fresh input + 3K output tokens
- GPT-5.4 mini
- $0.051
- Fits in one request
- Kimi K3
- $0.195
- Fits in one request
GPT-5.4 mini has the lower modeled cost
Costs use the listed standard API rates.
Cache-heavy agent loop
200K cached + 20K fresh input + 10K output tokens
- GPT-5.4 mini
- $0.075
- Fits in one request
- Kimi K3
- $0.27
- Fits in one request
GPT-5.4 mini 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.
SpecificationGPT-5.4 miniKimi K3
Context window
Maximum documented context; output-token limits may be lower.
GPT-5.4 mini
Kimi K3
1.05M
API model ID
GPT-5.4 mini
gpt-5.4-mini
OpenAI GPT-5.4 mini model documentationKimi K3
Not sourced
Cached-input rate
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
GPT-5.4 mini
$0.075 per 1M cached input tokens
OpenAI pricingKimi K3
$0.3 per 1M cached input tokens
Documented inputs
GPT-5.4 mini
text, image
OpenAI model catalogKimi K3
Not sourced
Documented outputs
GPT-5.4 mini
Kimi K3
Not sourced
Provider availability
GPT-5.4 mini
Generally Available · OpenAI Responses API
OpenAI model catalogKimi K3
Not sourced
Reasoning profile
GPT-5.4 mini
Reasoning
Kimi K3
Reasoning
Weight access
GPT-5.4 mini
Proprietary
Kimi K3
Pending
License
GPT-5.4 mini
Proprietary
Kimi K3
Pending
Release date
GPT-5.4 mini
2026-03-17
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 56.85, and the 90% score intervals do not overlap.
- Workload cost
- Repository review: $0.051 vs $0.195. Cache-heavy agent loop: $0.075 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 evidence47 rows
Agentic
Terminal-Bench 2.0
GPT-5.4 mini60%SourceKimi K388.3%SourceKimi K3 leads this result
OSWorld-Verified
GPT-5.4 mini72.1%SourceKimi K3—Not directly comparable
MCP Atlas
GPT-5.4 mini57.7%SourceKimi K384.2%SourceKimi K3 leads this result
Toolathlon
GPT-5.4 mini42.9%SourceKimi K3—Not directly comparable
τ²-bench results
GPT-5.4 mini93.4%SourceKimi K3—Not directly comparable
BrowseComp
GPT-5.4 mini—Kimi K391.2%SourceNot directly comparable
DeepSearchQA
GPT-5.4 mini—Kimi K395.0%SourceNot directly comparable
Toolathlon-Verified
GPT-5.4 mini—Kimi K373.2%SourceNot directly comparable
AutomationBench
GPT-5.4 mini—Kimi K330.8%SourceNot directly comparable
JobBench
GPT-5.4 mini—Kimi K352.9%SourceNot directly comparable
APEX-Agents
GPT-5.4 mini—Kimi K337.6%SourceNot directly comparable
SpreadsheetBench 2
GPT-5.4 mini—Kimi K334.8%SourceNot directly comparable
DECK-Bench
GPT-5.4 mini—Kimi K373.5%SourceNot directly comparable
Coding
Vibe Code Bench
GPT-5.4 mini47.97%SourceKimi K3—Not directly comparable
FrontierCode 1.1 Main
GPT-5.4 mini27.0%SourceKimi K3—Not directly comparable
EEBench
Shared sourceGPT-5.4 mini18.3%Kimi K338.3%Kimi K3 leads this result
deepSwe
GPT-5.4 mini—Kimi K367.5%SourceNot directly comparable
cursorBench32
GPT-5.4 mini—Kimi K360.8%SourceNot directly comparable
FrontierSWE
GPT-5.4 mini—Kimi K381.2%SourceNot directly comparable
ProgramBench
GPT-5.4 mini—Kimi K377.8%SourceNot directly comparable
Kimi Code Bench v2
GPT-5.4 mini—Kimi K372.9%SourceNot directly comparable
sweMarathon
GPT-5.4 mini—Kimi K342%SourceNot directly comparable
PostTrain Bench
GPT-5.4 mini—Kimi K336.6%SourceNot directly comparable
MLS-Bench Lite
GPT-5.4 mini—Kimi K348.3%SourceNot directly comparable
VulcanBench v3
GPT-5.4 mini—Kimi K373.9%SourceNot directly comparable
APEX-SWE
GPT-5.4 mini—Kimi K348.0%SourceNot directly comparable
InferenceEval
GPT-5.4 mini—Kimi K339.8%SourceNot directly comparable
KernelBench Internal
GPT-5.4 mini—Kimi K368.5%SourceNot directly comparable
Knowledge
Math
Multimodal
MMMU-Pro
GPT-5.4 mini76.6%SourceKimi K381.6%SourceKimi K3 leads this result
MMMU-Pro w/ Python
GPT-5.4 mini78%SourceKimi K383.4%SourceKimi K3 leads this result
OfficeQA Pro
GPT-5.4 mini—Kimi K363.3%SourceNot directly comparable
CharXiv w/o tools
GPT-5.4 mini—Kimi K384.8%SourceNot directly comparable
CharXiv
GPT-5.4 mini—Kimi K391.3%SourceNot directly comparable
MathVision
GPT-5.4 mini—Kimi K394.3%SourceNot directly comparable
MathVision w/ Python
GPT-5.4 mini—Kimi K397.8%SourceNot directly comparable
BabyVision w/ Python
GPT-5.4 mini—Kimi K385.7%SourceNot directly comparable
ZeroBench
GPT-5.4 mini—Kimi K323.0%SourceNot directly comparable
ZeroBench w/ Python
GPT-5.4 mini—Kimi K341.0%SourceNot directly comparable
WorldVQA ForceAnswer
GPT-5.4 mini—Kimi K351.0%SourceNot directly comparable
OmniDocBench
GPT-5.4 mini—Kimi K391.1%SourceNot directly comparable
PerceptionBench
GPT-5.4 mini—Kimi K358.5%SourceNot directly comparable
Frequently asked questions
Which is better, GPT-5.4 mini or Kimi K3?
Kimi K3 has the higher public score, 80.53 versus 56.85, 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, GPT-5.4 mini 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, GPT-5.4 mini 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, GPT-5.4 mini or Kimi K3?
For the stated presets, chat costs $0.003 on GPT-5.4 mini and $0.0105 on Kimi K3; repository review costs $0.051 and $0.195; the cache-heavy agent loop costs $0.075 and $0.27. Costs use the listed standard API rates.
Which has the larger context window, GPT-5.4 mini or Kimi K3?
Kimi K3 has the larger documented context window: 1.05M, compared with 400K.
Related comparisons
Last updated August 19, 2026
Watch GPT-5.4 mini 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.