Long documents
Prompts that approach the documented context limit
Gemini 3 Pro
Gemini 3 Pro has the larger documented context window.
Confidence: documented
Gemini 3 Pro 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.
Both of these models will change. Get the price, version and retirement notices for the pair, each with its source. Follow model changes
Share or export
Decision reading
Kimi K3 has the higher public score estimate, 72.14 versus 61.57, but the 90% score intervals overlap. Treat that as a lead, not a settled winner. 5 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.
- Shared results
- 5
- Gemini 3 Pro only
- 7
- Kimi K3 only
- 43
- Like-for-like categories
- 0 / 8
Supported: Gemini 3 Pro and Kimi K3How 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.
Chat turn cost
1K fresh input + 500 output tokens
Gemini 3 Pro
Gemini 3 Pro has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-ratesCache-heavy agent loop cost
200K cached + 20K fresh input + 10K output tokens
Kimi K3
Kimi K3 has the lower estimated token cost for this stated workload. Gemini 3 Pro has no published cached-input rate, so cached tokens use its listed input rate.
Confidence: rate-fallback
Show secondary and unsupported calls
Repository review cost
50K fresh input + 3K output tokens
Gemini 3 Pro
Gemini 3 Pro has the lower estimated token cost for this stated workload. Costs use the listed standard API rates.
Confidence: listed-ratesCoding work
Code generation, repair, and software-engineering tasks
Not enough matched evidence
Gemini 3 Pro is scored on Estimated evidence for coding, so the reading is directional rather than like-for-like.
Confidence: limitedAgentic work
Tool use, computer use, and multi-step task completion
Not enough matched evidence
Gemini 3 Pro is not ranked on the public lane for agentic, so no winner is named for agentic.
Confidence: limited
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.
The same task catalog as the LLM Selector. Each job names the evidence surface it rests on; nothing here adds a new score.
45.4Gemini 3 Pro61.4Kimi K3
Directional only · 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.
Same basis rules as the category table below
What is actually comparable
Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.
7 Gemini 3 Pro only
5 Shared
43 Kimi K3 only
3 categories rest on Estimated evidence or different evidence sets. Those rows remain visible for coverage context but do not name a winner.
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.
ARC-AGI-2Reasoning
Normalized gap 29.3Gemini 3 Pro:31.1%Kimi K3:60.4%CharXivMultimodal
Normalized gap 9.9Gemini 3 Pro:81.4%Kimi K3:91.3%MMMU-ProMultimodal
Normalized gap 0.6Gemini 3 Pro:81%Kimi K3:81.6%
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.
Coding
Directional only- Gemini 3 Pro
- 45.4
- Estimated · #53/144
- Kimi K3
- 61.4
- Supported · #18/144
- Basis
- BenchAlign v5.8 lane · 1 vs 14 public rows
- Reading
- Directional only
Multimodal
Directional only- Gemini 3 Pro
- 75.2
- #17/49
- Kimi K3
- 89.4
- #1/49
- Basis
- Provisional lane · 2 vs 3 weighted rows
- Reading
- Directional only
Knowledge
Directional only- Gemini 3 Pro
- 59.2
- Estimated · #43/171
- Kimi K3
- 67.9
- Supported · #18/171
- Basis
- BenchAlign v5.8 lane · 0 vs 6 public rows
- Reading
- Directional only
Agentic
Not comparable- Gemini 3 Pro
- Not ranked
- Kimi K3
- 68.1
- Supported · #8/119
- Basis
- BenchAlign v5.8 lane · 2 vs 12 public rows
- Reading
- Not comparable
Reasoning
Not comparable- Gemini 3 Pro
- 46.5
- Unranked · 3 rankable rows
- Kimi K3
- 65.8
- #18/27
- Basis
- Provisional lane · 1 vs 1 weighted rows
- Reading
- Not comparable
Multilingual
Not comparable- Gemini 3 Pro
- Not ranked
- Kimi K3
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Instruction following
Not comparable- Gemini 3 Pro
- 84.7
- #38/125
- Kimi K3
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Math
Not comparable- Gemini 3 Pro
- 55.2
- Unranked · 2 rankable rows
- Kimi K3
- Not ranked
- Basis
- Provisional lane · 2 vs 0 weighted rows
- Reading
- Not comparable
| Category | Gemini 3 Pro | Kimi K3 | Basis | Reading |
|---|---|---|---|---|
| Coding | 45.4Estimated · #53/144 | 61.4Supported · #18/144 | Directional onlyBenchAlign v5.8 lane · 1 vs 14 public rows | Directional only |
| Multimodal | 75.2#17/49 | 89.4#1/49 | Directional onlyProvisional lane · 2 vs 3 weighted rows | Directional only |
| Knowledge | 59.2Estimated · #43/171 | 67.9Supported · #18/171 | Directional onlyBenchAlign v5.8 lane · 0 vs 6 public rows | Directional only |
| Agentic | Not ranked | 68.1Supported · #8/119 | Not comparableBenchAlign v5.8 lane · 2 vs 12 public rows | Not comparable |
| Reasoning | 46.5Unranked · 3 rankable rows | 65.8#18/27 | Not comparableProvisional lane · 1 vs 1 weighted rows | Not comparable |
| Multilingual | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Instruction following | 84.7#38/125 | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Math | 55.2Unranked · 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, 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
- Gemini 3 Pro
- $0.008
- Fits in one request
- Kimi K3
- $0.0105
- Fits in one request
Gemini 3 Pro has the lower modeled cost
Costs use the listed standard API rates.
Repository review
50K fresh input + 3K output tokens
- Gemini 3 Pro
- $0.136
- Fits in one request
- Kimi K3
- $0.195
- Fits in one request
Gemini 3 Pro has the lower modeled cost
Costs use the listed standard API rates.
Cache-heavy agent loop
200K cached + 20K fresh input + 10K output tokens
- Gemini 3 Pro
- $0.56
- Fits in one request
- Cached input priced at the published list-input rate
- Kimi K3
- $0.27
- Fits in one request
Kimi K3 has the lower modeled cost
Gemini 3 Pro has no published cached-input rate, so cached tokens use its listed input rate.
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.
SpecificationGemini 3 ProKimi K3
Context window
Maximum documented context; output-token limits may be lower.
Gemini 3 Pro
2M
Kimi K3
1.05M
API model ID
Gemini 3 Pro
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.
Gemini 3 Pro
Not published
Kimi K3
$0.3 per 1M cached input tokens
Documented inputs
Gemini 3 Pro
Not sourced
Kimi K3
Not sourced
Documented outputs
Gemini 3 Pro
Not sourced
Kimi K3
Not sourced
Provider availability
Gemini 3 Pro
Not sourced
Kimi K3
Not sourced
Reasoning profile
Gemini 3 Pro
Non-Reasoning
Kimi K3
Reasoning
Weight access
Gemini 3 Pro
Proprietary
Kimi K3
Pending
License
Gemini 3 Pro
Proprietary
Kimi K3
Pending
Release date
Gemini 3 Pro
2025-11-18
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, 72.14 versus 61.57, but the 90% score intervals overlap.
- Workload cost
- Repository review: $0.136 vs $0.195. Cache-heavy agent loop: $0.56 vs $0.27.
- Context tradeoff
- Gemini 3 Pro has the larger documented window (2M).
Run the same representative tasks against both endpoints before changing production traffic.
Questions
Which is better, Gemini 3 Pro or Kimi K3?
Kimi K3 has the higher public score estimate, 72.14 versus 61.57, but the 90% score intervals overlap. The higher estimate is not a decisive winner because the uncertainty ranges overlap.
Which is better for coding, Gemini 3 Pro or Kimi K3?
Kimi K3 scores higher for coding on the public lane, 61.4 to 45.4. Gemini 3 Pro is scored on Estimated evidence for coding, so the reading is directional rather than like-for-like. BenchLM does not name a winner for coding from a directional reading; read the shared benchmark rows directly and test the models on the same task set.
Which is better for agentic tasks, Gemini 3 Pro or Kimi K3?
Gemini 3 Pro is not ranked on the public lane for agentic tasks, so no winner is named for agentic tasks.
Which costs less, Gemini 3 Pro or Kimi K3?
For the stated presets, chat costs $0.008 on Gemini 3 Pro and $0.0105 on Kimi K3; repository review costs $0.136 and $0.195; the cache-heavy agent loop costs $0.56 and $0.27. Gemini 3 Pro has no published cached-input rate, so cached tokens use its listed input rate.
Which has the larger context window, Gemini 3 Pro or Kimi K3?
Gemini 3 Pro has the larger documented context window: 2M, compared with 1.05M.
Benchmark evidence
The full public result ledger is available for audit without forcing a wide desktop table onto a phone.
Browse raw public benchmark evidence55 rows
Agentic
Gert Labs
Gemini 3 Pro63.23%SourceKimi K3—Not directly comparable
JobBench
Gemini 3 Pro11.4%SourceKimi K352.9%SourceKimi K3 leads this result
Terminal-Bench 2.1
Gemini 3 Pro—Kimi K388.3%SourceNot directly comparable
BrowseComp
Gemini 3 Pro—Kimi K391.2%SourceNot directly comparable
DeepSearchQA
Gemini 3 Pro—Kimi K395.0%SourceNot directly comparable
Toolathlon-Verified
Gemini 3 Pro—Kimi K373.2%SourceNot directly comparable
MCP Atlas
Gemini 3 Pro—Kimi K384.2%SourceNot directly comparable
AutomationBench
Gemini 3 Pro—Kimi K330.8%SourceNot directly comparable
APEX-Agents
Gemini 3 Pro—Kimi K337.6%SourceNot directly comparable
SpreadsheetBench 2
Gemini 3 Pro—Kimi K334.8%SourceNot directly comparable
DECK-Bench
Gemini 3 Pro—Kimi K373.5%SourceNot directly comparable
Terminal-Bench 2.1 (Vals)
Gemini 3 Pro—Kimi K380.9%SourceNot directly comparable
ApprenticeBench
Gemini 3 Pro—Kimi K318%SourceNot directly comparable
Coding
Vibe Code Bench
Gemini 3 Pro14.30%SourceKimi K3—Not directly comparable
DeepSWE
Gemini 3 Pro—Kimi K367.5%SourceNot directly comparable
CursorBench 3.2
Gemini 3 Pro—Kimi K360.8%SourceNot directly comparable
FrontierSWE
Gemini 3 Pro—Kimi K381.2%SourceNot directly comparable
ProgramBench
Gemini 3 Pro—Kimi K377.8%SourceNot directly comparable
Kimi Code Bench v2
Gemini 3 Pro—Kimi K372.9%SourceNot directly comparable
sweMarathon
Gemini 3 Pro—Kimi K342%SourceNot directly comparable
PostTrain Bench
Gemini 3 Pro—Kimi K336.6%SourceNot directly comparable
MLS-Bench Lite
Gemini 3 Pro—Kimi K348.3%SourceNot directly comparable
VulcanBench v3
Gemini 3 Pro—Kimi K373.7%SourceNot directly comparable
OpenHarmony Bench
Gemini 3 Pro—Kimi K357.3%SourceNot directly comparable
FrontierSWE v2
Gemini 3 Pro—Kimi K325.9%SourceNot directly comparable
LiveCodeBench (Vals)
Gemini 3 Pro—Kimi K387.2%SourceNot directly comparable
SWE-bench (Vals)
Gemini 3 Pro—Kimi K393.4%SourceNot directly comparable
PostTrainBench v1.1
Gemini 3 Pro—Kimi K332.0%SourceNot directly comparable
Reasoning
Multimodal
MMMU-Pro
Gemini 3 Pro81%SourceKimi K381.6%SourceKimi K3 leads this result
MathVision
Gemini 3 Pro86.6%SourceKimi K394.3%SourceKimi K3 leads this result
VideoMMMU
Gemini 3 Pro87.6%SourceKimi K3—Not directly comparable
ScreenSpot Pro
Gemini 3 Pro72.7%SourceKimi K3—Not directly comparable
CharXiv
Gemini 3 Pro81.4%SourceKimi K391.3%SourceKimi K3 leads this result
V*
Gemini 3 Pro88.0%SourceKimi K3—Not directly comparable
OfficeQA Pro
Gemini 3 Pro—Kimi K363.3%SourceNot directly comparable
MMMU-Pro w/ Python
Gemini 3 Pro—Kimi K383.4%SourceNot directly comparable
CharXiv w/o tools
Gemini 3 Pro—Kimi K384.8%SourceNot directly comparable
MathVision w/ Python
Gemini 3 Pro—Kimi K397.8%SourceNot directly comparable
BabyVision w/ Python
Gemini 3 Pro—Kimi K385.7%SourceNot directly comparable
ZeroBench
Gemini 3 Pro—Kimi K323.0%SourceNot directly comparable
ZeroBench w/ Python
Gemini 3 Pro—Kimi K341.0%SourceNot directly comparable
WorldVQA ForceAnswer
Gemini 3 Pro—Kimi K351.0%SourceNot directly comparable
OmniDocBench
Gemini 3 Pro—Kimi K391.1%SourceNot directly comparable
PerceptionBench
Gemini 3 Pro—Kimi K358.5%SourceNot directly comparable
Knowledge
GPQA
Gemini 3 Pro—Kimi K393.5%SourceNot directly comparable
GPQA-D
Gemini 3 Pro—Kimi K393.5%SourceNot directly comparable
HLE
Gemini 3 Pro—Kimi K356%SourceNot directly comparable
HLE w/o tools
Gemini 3 Pro—Kimi K343.5%SourceNot directly comparable
GPQA Diamond (Vals)
Gemini 3 Pro—Kimi K392.9%SourceNot directly comparable
MMLU-Pro (Vals)
Gemini 3 Pro—Kimi K388.0%SourceNot directly comparable
Instruction following
Gray Swan IPI (15 attempts)
Gemini 3 Pro—Kimi K352.7%SourceNot directly comparable
Math
55 public results · 5 shared
Watch Gemini 3 Pro 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.
Last updated October 2, 2026