Long documents
Prompts that approach the documented context limit
Kimi K2.6
Kimi K2.6 has the larger documented context window.
Confidence: documented
Grok 3 Mini vs Kimi K2.6
Updated October 26, 2026. Rank says Kimi K2.6 is ahead. Price, access, and your workload can each overturn that. Public scores include evidence status and uncertainty.
Share or export
Decision reading
Kimi K2.6 has the higher public point estimate, 58.36 versus 44.28. Their conditional score ranges overlap. These ranges do not establish rank confidence. 1 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.
- Shared results
- 1
- Grok 3 Mini only
- 2
- Kimi K2.6 only
- 36
- Like-for-like categories
- 0 / 8
Estimated: Grok 3 Mini and Kimi K2.6. Conditional ranges do not establish rank confidence.How 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
Grok 3 Mini
Grok 3 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
Grok 3 Mini
Grok 3 Mini 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
Grok 3 Mini is not ranked on the public lane for coding, so no winner is named for coding.
Confidence: limitedAgentic work
Tool use, computer use, and multi-step task completion
Not enough matched evidence
Grok 3 Mini is not ranked on the public lane for agentic, so no winner is named for agentic.
Confidence: limitedCache-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. Grok 3 Mini does not fit this workload in one request. Grok 3 Mini has no published cached-input rate, so cached tokens use its listed input rate.
Confidence: rate-fallback
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.
—Grok 3 Mini43.5Kimi K2.6
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.
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.
2 Grok 3 Mini only
1 Shared
36 Kimi K2.6 only
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.
GPQAKnowledge
Normalized gap 10.2Grok 3 Mini:80.3%Kimi K2.6:90.5%
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.
Agentic
Not comparable- Grok 3 Mini
- Not ranked
- Kimi K2.6
- 44.1
- Supported · #64/133
- Basis
- BenchAlign v5.8 lane · 0 vs 12 public rows
- Reading
- Not comparable
Coding
Not comparable- Grok 3 Mini
- Not ranked
- Kimi K2.6
- 43.5
- Supported · #61/154
- Basis
- BenchAlign v5.8 lane · 1 vs 10 public rows
- Reading
- Not comparable
Reasoning
Not comparable- Grok 3 Mini
- Not ranked
- Kimi K2.6
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Multimodal
Not comparable- Grok 3 Mini
- Not ranked
- Kimi K2.6
- 73.5
- #30/58
- Basis
- Provisional lane · 0 vs 2 weighted rows
- Reading
- Not comparable
Knowledge
Not comparable- Grok 3 Mini
- Not ranked
- Kimi K2.6
- 59.5
- Supported · #50/188
- Basis
- BenchAlign v5.8 lane · 1 vs 5 public rows
- Reading
- Not comparable
Multilingual
Not comparable- Grok 3 Mini
- Not ranked
- Kimi K2.6
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Instruction following
Not comparable- Grok 3 Mini
- Not ranked
- Kimi K2.6
- 83.5
- #57/155
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Math
Not comparable- Grok 3 Mini
- Not ranked
- Kimi K2.6
- 76.8
- #4/12
- Basis
- Provisional lane · 0 vs 4 weighted rows
- Reading
- Not comparable
| Category | Grok 3 Mini | Kimi K2.6 | Basis | Reading |
|---|---|---|---|---|
| Agentic | Not ranked | 44.1Supported · #64/133 | Not comparableBenchAlign v5.8 lane · 0 vs 12 public rows | Not comparable |
| Coding | Not ranked | 43.5Supported · #61/154 | Not comparableBenchAlign v5.8 lane · 1 vs 10 public rows | Not comparable |
| Reasoning | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Multimodal | Not ranked | 73.5#30/58 | Not comparableProvisional lane · 0 vs 2 weighted rows | Not comparable |
| Knowledge | Not ranked | 59.5Supported · #50/188 | Not comparableBenchAlign v5.8 lane · 1 vs 5 public rows | Not comparable |
| Multilingual | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Instruction following | Not ranked | 83.5#57/155 | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Math | Not ranked | 76.8#4/12 | Not comparableProvisional lane · 0 vs 4 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
- Grok 3 Mini
- $0.00055
- Fits in one request
- Kimi K2.6
- $0.00295
- Fits in one request
Grok 3 Mini has the lower modeled cost
Costs use the listed standard API rates.
Repository review
50K fresh input + 3K output tokens
- Grok 3 Mini
- $0.0165
- Fits in one request
- Kimi K2.6
- $0.0595
- Fits in one request
Grok 3 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
- Grok 3 Mini
- $0.071
- Does not fit in one request
- Cached input priced at the published list-input rate
- Kimi K2.6
- $0.091
- Fits in one request
Grok 3 Mini does not fit this workload in one request. Grok 3 Mini 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.
SpecificationGrok 3 MiniKimi K2.6
Context window
Maximum documented context; output-token limits may be lower.
Grok 3 Mini
128K
Kimi K2.6
256K
API model ID
Grok 3 Mini
Not sourced
Kimi K2.6
Not sourced
Cached-input rate
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
Grok 3 Mini
Not published
Kimi K2.6
$0.16 per 1M cached input tokens
Documented inputs
Grok 3 Mini
Not sourced
Kimi K2.6
Not sourced
Documented outputs
Grok 3 Mini
Not sourced
Kimi K2.6
Not sourced
Provider availability
Grok 3 Mini
Not sourced
Kimi K2.6
Not sourced
Reasoning profile
Grok 3 Mini
Reasoning
Kimi K2.6
Reasoning
Weight access
Grok 3 Mini
Proprietary
Kimi K2.6
Open Weight
License
Grok 3 Mini
Proprietary
Kimi K2.6
Open Weight
Release date
Grok 3 Mini
2025-02-19
Kimi K2.6
2026-04-20
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 K2.6 has the higher public point estimate, 58.36 versus 44.28. Their conditional score ranges overlap. These ranges do not establish rank confidence.
- Workload cost
- Repository review: $0.0165 vs $0.0595. Cache-heavy agent loop: $0.071 vs $0.091.
- Context tradeoff
- Kimi K2.6 has the larger documented window (256K).
Run the same representative tasks against both endpoints before changing production traffic.
Questions
Which is better, Grok 3 Mini or Kimi K2.6?
Kimi K2.6 has the higher public point estimate, 58.36 versus 44.28. Their conditional score ranges overlap. These ranges do not establish rank confidence. The page therefore keeps the decision tied to the specific documented workload.
Which is better for coding, Grok 3 Mini or Kimi K2.6?
Grok 3 Mini is not ranked on the public lane for coding, so no winner is named for coding.
Which is better for agentic tasks, Grok 3 Mini or Kimi K2.6?
Grok 3 Mini is not ranked on the public lane for agentic tasks, so no winner is named for agentic tasks.
Which costs less, Grok 3 Mini or Kimi K2.6?
For the stated presets, chat costs $0.00055 on Grok 3 Mini and $0.00295 on Kimi K2.6; repository review costs $0.0165 and $0.0595; the cache-heavy agent loop costs $0.071 and $0.091. Grok 3 Mini does not fit this workload in one request. Grok 3 Mini has no published cached-input rate, so cached tokens use its listed input rate.
Which has the larger context window, Grok 3 Mini or Kimi K2.6?
Kimi K2.6 has the larger documented context window: 256K, compared with 128K.
Self-host vs API cost
Estimates at 50,000 req/day · 1000 tokens/req average.
Grok 3 Mini
API / mo$600
Self-host / moNot listed
Break-even—
Proprietary model — self-hosting not applicable.
Kimi K2.6
API / mo$3,713
Self-host / mo$27,962
Break-even500M/day
Benchmark evidence
The full public result ledger is available for audit without forcing a wide desktop table onto a phone.
Browse raw public benchmark evidence39 rows
Agentic
Terminal-Bench 2.0
Grok 3 Mini—Kimi K2.666.7%SourceNot directly comparable
BrowseComp
Grok 3 Mini—Kimi K2.683.2%SourceNot directly comparable
OSWorld-Verified
Grok 3 Mini—Kimi K2.673.1%SourceNot directly comparable
Toolathlon
Grok 3 Mini—Kimi K2.650%SourceNot directly comparable
MCP Atlas
Grok 3 Mini—Kimi K2.655.9%SourceNot directly comparable
Claw-Eval
Grok 3 Mini—Kimi K2.662.3%SourceNot directly comparable
DeepSearchQA
Grok 3 Mini—Kimi K2.692.5%SourceNot directly comparable
WideResearch
Grok 3 Mini—Kimi K2.680.8%SourceNot directly comparable
Gert Labs
Grok 3 Mini—Kimi K2.656.82%SourceNot directly comparable
ResearchClawBench
Grok 3 Mini—Kimi K2.618.0%SourceNot directly comparable
OSWorld 2.0
Grok 3 Mini—Kimi K2.64.6%SourceNot directly comparable
Terminal-Bench 2.1 (Vals)
Grok 3 Mini—Kimi K2.653.6%SourceNot directly comparable
Coding
LiveCodeBench
Grok 3 Mini74.8%SourceKimi K2.6—Not directly comparable
SWE-bench Verified
Grok 3 Mini—Kimi K2.680.2%SourceNot directly comparable
LiveCodeBench v6
Grok 3 Mini—Kimi K2.689.6%SourceNot directly comparable
SWE-bench Pro
Grok 3 Mini—Kimi K2.658.6%SourceNot directly comparable
SWE Multilingual
Grok 3 Mini—Kimi K2.676.7%SourceNot directly comparable
SciCode
Grok 3 Mini—Kimi K2.652.2%SourceNot directly comparable
Terminal-Bench 2.0
Grok 3 Mini—Kimi K2.666.7%SourceNot directly comparable
Vibe Code Bench
Grok 3 Mini—Kimi K2.637.89%SourceNot directly comparable
cursorBench31
Grok 3 Mini—Kimi K2.647.6%SourceNot directly comparable
LiveCodeBench (Vals)
Grok 3 Mini—Kimi K2.686.8%SourceNot directly comparable
SWE-bench (Vals)
Grok 3 Mini—Kimi K2.676.2%SourceNot directly comparable
Multimodal
MMMU-Pro
Grok 3 Mini—Kimi K2.679.4%SourceNot directly comparable
MMMU-Pro w/ Python
Grok 3 Mini—Kimi K2.680.1%SourceNot directly comparable
CharXiv
Grok 3 Mini—Kimi K2.680.4%SourceNot directly comparable
MathVision
Grok 3 Mini—Kimi K2.687.4%SourceNot directly comparable
V*
Grok 3 Mini—Kimi K2.696.9%SourceNot directly comparable
Knowledge
GPQA
Grok 3 Mini80.3%SourceKimi K2.690.5%SourceKimi K2.6 leads this result
GPQA-D
Grok 3 Mini—Kimi K2.690.5%SourceNot directly comparable
HLE
Grok 3 Mini—Kimi K2.634.7%SourceNot directly comparable
GPQA Diamond (Vals)
Grok 3 Mini—Kimi K2.689.1%SourceNot directly comparable
MMLU-Pro (Vals)
Grok 3 Mini—Kimi K2.687.6%SourceNot directly comparable
Math
AIME 2024
Grok 3 Mini89.5%SourceKimi K2.6—Not directly comparable
AIME26
Grok 3 Mini—Kimi K2.696.4%SourceNot directly comparable
HMMT Feb 2026
Grok 3 Mini—Kimi K2.692.7%SourceNot directly comparable
MMAnswerBench
Grok 3 Mini—Kimi K2.686.0%SourceNot directly comparable
FrontierMath v2 (Tiers 1-3)
Grok 3 Mini—Kimi K2.657.193%SourceNot directly comparable
FrontierMath v2 (Tier 4)
Grok 3 Mini—Kimi K2.625.641%SourceNot directly comparable
39 public results · 1 shared
Watch Grok 3 Mini vs Kimi K2.6
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 6,100+ subscribers.
Last updated October 26, 2026