Long documents
Prompts that approach the documented context limit
Kimi K2.5
Kimi K2.5 has the larger documented context window.
Confidence: documented
Grok 4 vs Kimi K2.5
Updated September 22, 2026. Rank says Kimi K2.5 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 K2.5 has the higher public score estimate, 53.79 versus 52.71, but the 90% score intervals overlap. Treat that as a lead, not a settled winner. 4 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.
- Shared results
- 4
- Grok 4 only
- 0
- Kimi K2.5 only
- 41
- Like-for-like categories
- 0 / 8
Supported: Grok 4 and Kimi K2.5How 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.
Show secondary and unsupported calls
Coding work
Code generation, repair, and software-engineering tasks
Not enough matched evidence
Grok 4 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 4 is not ranked on the public lane for agentic, so no winner is named for agentic.
Confidence: limitedChat turn cost
1K fresh input + 500 output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
Confidence: listed-ratesCache-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 4 does not fit this workload in one request. Kimi K2.5 has no published cached-input rate, so cached tokens use its listed input rate. Grok 4 has no comparable published API token rate.
Confidence: rate-fallbackRepository review cost
50K fresh input + 3K output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
Confidence: listed-rates
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 440.9Kimi K2.5
Not comparable · BenchAlign v5.6
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.
4 Shared
41 Kimi K2.5 only
2 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.
FrontierMath v2 (Tiers 1-3)Math
Normalized gap 8.2Grok 4:19.655%Kimi K2.5:27.900%FrontierMath v2 (Tier 4)Math
Normalized gap 2.1Grok 4:2.083%Kimi K2.5:4.200%
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 v5.6 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.
Knowledge
Directional only- Grok 4
- 48.1
- Estimated · #61/160
- Kimi K2.5
- 46.4
- Estimated · #68/160
- Basis
- BenchAlign v5.6 lane · 0 vs 6 public rows
- Reading
- Directional only
Instruction following
Directional only- Grok 4
- 62.9
- #71/124
- Kimi K2.5
- 84.4
- #41/124
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Directional only
Agentic
Not comparable- Grok 4
- Not ranked
- Kimi K2.5
- 35.5
- Estimated · #54/105
- Basis
- BenchAlign v5.6 lane · 1 vs 14 public rows
- Reading
- Not comparable
Coding
Not comparable- Grok 4
- Not ranked
- Kimi K2.5
- 40.9
- Estimated · #58/135
- Basis
- BenchAlign v5.6 lane · 1 vs 8 public rows
- Reading
- Not comparable
Reasoning
Not comparable- Grok 4
- 67.7
- Unranked · 2 rankable rows
- Kimi K2.5
- 54.1
- Unranked · 3 rankable rows
- Basis
- Provisional lane · 0 vs 1 weighted rows
- Reading
- Not comparable
Multimodal
Not comparable- Grok 4
- 61.4
- Unranked · 1 rankable row
- Kimi K2.5
- 65.8
- #25/50
- Basis
- Provisional lane · 0 vs 1 weighted rows
- Reading
- Not comparable
Multilingual
Not comparable- Grok 4
- Not ranked
- Kimi K2.5
- 38.2
- #8/12
- Basis
- Provisional lane · 0 vs 1 weighted rows
- Reading
- Not comparable
Math
Not comparable- Grok 4
- 38.3
- Unranked · 2 rankable rows
- Kimi K2.5
- 62.1
- #4/7
- Basis
- Provisional lane · 2 vs 4 weighted rows
- Reading
- Not comparable
| Category | Grok 4 | Kimi K2.5 | Basis | Reading |
|---|---|---|---|---|
| Knowledge | 48.1Estimated · #61/160 | 46.4Estimated · #68/160 | Directional onlyBenchAlign v5.6 lane · 0 vs 6 public rows | Directional only |
| Instruction following | 62.9#71/124 | 84.4#41/124 | Directional onlyProvisional lane · 0 vs 0 weighted rows | Directional only |
| Agentic | Not ranked | 35.5Estimated · #54/105 | Not comparableBenchAlign v5.6 lane · 1 vs 14 public rows | Not comparable |
| Coding | Not ranked | 40.9Estimated · #58/135 | Not comparableBenchAlign v5.6 lane · 1 vs 8 public rows | Not comparable |
| Reasoning | 67.7Unranked · 2 rankable rows | 54.1Unranked · 3 rankable rows | Not comparableProvisional lane · 0 vs 1 weighted rows | Not comparable |
| Multimodal | 61.4Unranked · 1 rankable row | 65.8#25/50 | Not comparableProvisional lane · 0 vs 1 weighted rows | Not comparable |
| Multilingual | Not ranked | 38.2#8/12 | Not comparableProvisional lane · 0 vs 1 weighted rows | Not comparable |
| Math | 38.3Unranked · 2 rankable rows | 62.1#4/7 | Not comparableProvisional lane · 2 vs 4 weighted rows | Not comparable |
Ranks count the models scored in each category’s lane, so the agentic, coding, and knowledge denominators (BenchAlign v5.6) 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.
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 4
- API rate not published
- Fits in one request
- Kimi K2.5
- $0.0021
- Fits in one request
Grok 4 has no comparable published API token rate.
Repository review
50K fresh input + 3K output tokens
- Grok 4
- API rate not published
- Fits in one request
- Kimi K2.5
- $0.039
- Fits in one request
Grok 4 has no comparable published API token rate.
Cache-heavy agent loop
200K cached + 20K fresh input + 10K output tokens
- Grok 4
- API rate not published
- Does not fit in one request
- Cached-input rate unavailable
- Kimi K2.5
- $0.162
- Fits in one request
- Cached input priced at the published list-input rate
Grok 4 does not fit this workload in one request. Kimi K2.5 has no published cached-input rate, so cached tokens use its listed input rate. Grok 4 has no comparable published API token 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 4Kimi K2.5
Context window
Maximum documented context; output-token limits may be lower.
Grok 4
128K
Kimi K2.5
256K
API model ID
Grok 4
Not sourced
Kimi K2.5
Not sourced
Cached-input rate
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
Grok 4
No comparable hosted API rate
Kimi K2.5
Not published
Documented inputs
Grok 4
Not sourced
Kimi K2.5
Not sourced
Documented outputs
Grok 4
Not sourced
Kimi K2.5
Not sourced
Provider availability
Grok 4
Not sourced
Kimi K2.5
Not sourced
Reasoning profile
Grok 4
Non-Reasoning
Kimi K2.5
Non-Reasoning
Weight access
Grok 4
Proprietary
Kimi K2.5
Open Weight
License
Grok 4
Proprietary
Kimi K2.5
Open Weight
Release date
Grok 4
2025-07-09
Kimi K2.5
2026-02-01
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 K2.5 has the higher public score estimate, 53.79 versus 52.71, but the 90% score intervals overlap.
- Workload cost
- A complete comparable API-rate estimate is not available for both models.
- Context tradeoff
- Kimi K2.5 has the larger documented window (256K).
Run the same representative tasks against both endpoints before changing production traffic.
Questions
Which is better, Grok 4 or Kimi K2.5?
Kimi K2.5 has the higher public score estimate, 53.79 versus 52.71, but the 90% score intervals overlap. The higher estimate is not a decisive winner because the uncertainty ranges overlap.
Which is better for coding, Grok 4 or Kimi K2.5?
Grok 4 is not ranked on the public lane for coding, so no winner is named for coding.
Which is better for agentic tasks, Grok 4 or Kimi K2.5?
Grok 4 is not ranked on the public lane for agentic tasks, so no winner is named for agentic tasks.
Which costs less, Grok 4 or Kimi K2.5?
Both models do not have comparable published API token rates, so this page does not name a universal price winner.
Which has the larger context window, Grok 4 or Kimi K2.5?
Kimi K2.5 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 4
API / mo$0
Self-host / moNot listed
Break-even—
Proprietary model — self-hosting not applicable.
Kimi K2.5
API / mo$2,700
Self-host / mo$5,221
Break-even132M/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 evidence45 rows
Agentic
Gert Labs
Shared sourceGrok 442.34%Kimi K2.545.88%Kimi K2.5 leads this result
Terminal-Bench 2.0
Grok 4—Kimi K2.550.8%SourceNot directly comparable
BrowseComp
Grok 4—Kimi K2.560.6%SourceNot directly comparable
Claw-Eval
Grok 4—Kimi K2.552.3%SourceNot directly comparable
QwenClawBench
Grok 4—Kimi K2.554.3%SourceNot directly comparable
τ³-bench results
Grok 4—Kimi K2.565.7%SourceNot directly comparable
DeepSearchQA
Grok 4—Kimi K2.577.1%SourceNot directly comparable
DeepPlanning
Grok 4—Kimi K2.514.4%SourceNot directly comparable
Toolathlon
Grok 4—Kimi K2.527.8%SourceNot directly comparable
MCP Atlas
Grok 4—Kimi K2.529.5%SourceNot directly comparable
MCP-Tasks
Grok 4—Kimi K2.559.1%SourceNot directly comparable
WideResearch
Grok 4—Kimi K2.572.7%SourceNot directly comparable
ResearchClawBench
Grok 4—Kimi K2.514.0%SourceNot directly comparable
JobBench
Grok 4—Kimi K2.58.7%SourceNot directly comparable
Coding
React Native Evals
Shared sourceGrok 472.6%Kimi K2.577.2%Kimi K2.5 leads this result
SWE-bench Verified
Grok 4—Kimi K2.576.8%SourceNot directly comparable
SWE-bench Verified*
Grok 4—Kimi K2.570.8%SourceNot directly comparable
LiveCodeBench v6
Grok 4—Kimi K2.585.0%SourceNot directly comparable
SWE-bench Pro
Grok 4—Kimi K2.550.7%SourceNot directly comparable
SWE Multilingual
Grok 4—Kimi K2.573%SourceNot directly comparable
SWE-Rebench
Grok 4—Kimi K2.558.5%SourceNot directly comparable
SciCode
Grok 4—Kimi K2.548.7%SourceNot directly comparable
Reasoning
LongBench v2
Grok 4—Kimi K2.561%SourceNot directly comparable
Multimodal
Knowledge
GPQA
Grok 4—Kimi K2.587.6%SourceNot directly comparable
GPQA-D
Grok 4—Kimi K2.587.6%SourceNot directly comparable
SuperGPQA
Grok 4—Kimi K2.569.2%SourceNot directly comparable
MMLU-Pro
Grok 4—Kimi K2.587.1%SourceNot directly comparable
MMLU-Pro (Arcee)
Grok 4—Kimi K2.587.1%SourceNot directly comparable
HLE
Grok 4—Kimi K2.530.1%SourceNot directly comparable
Multilingual
Instruction following
IFEval
Grok 4—Kimi K2.593.9%SourceNot directly comparable
Math
FrontierMath v2 (Tiers 1-3)
Shared sourceGrok 419.655%Kimi K2.527.900%Kimi K2.5 leads this result
FrontierMath v2 (Tier 4)
Shared sourceGrok 42.083%Kimi K2.54.200%Kimi K2.5 leads this result
AIME 2025
Grok 4—Kimi K2.596.1%SourceNot directly comparable
AIME26
Grok 4—Kimi K2.595.8%SourceNot directly comparable
AIME25 (Arcee)
Grok 4—Kimi K2.596.3%SourceNot directly comparable
HMMT Feb 2025
Grok 4—Kimi K2.595.4%SourceNot directly comparable
HMMT Nov 2025
Grok 4—Kimi K2.591.1%SourceNot directly comparable
HMMT Feb 2026
Grok 4—Kimi K2.587.1%SourceNot directly comparable
MMAnswerBench
Grok 4—Kimi K2.581.8%SourceNot directly comparable
45 public results · 4 shared
Watch Grok 4 vs Kimi K2.5
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 September 22, 2026