Agentic
Directional only- Hy3 Preview
- 54.4
- Kimi K3
- 89.5
- Weighted basis
- 1 vs 2 rows
- Reading
- Directional only
Hy3 Preview 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 43.9, and the 90% score intervals do not overlap.
3 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
Show secondary and unsupported calls
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
Chat 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-rates
Cache-heavy agent loop cost
200K cached + 20K fresh input + 10K output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
Confidence: listed-rates
Repository 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
What is actually comparable
Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.
3 Hy3 Preview only
3 Shared
37 Kimi K3 only
- Shared results
- 3
- Hy3 Preview only
- 3
- Kimi K3 only
- 37
- Like-for-like categories
- 0 / 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.
Knowledge
Directional only- Hy3 Preview
- 87.2
- Kimi K3
- 61.0
- Weighted basis
- 1 vs 2 rows
- Reading
- Directional only
Coding
Not comparable- Hy3 Preview
- 74.4
- Kimi K3
- Not measured
- Weighted basis
- 1 vs 0 rows
- Reading
- Not comparable
Reasoning
Not comparable- Hy3 Preview
- Not measured
- Kimi K3
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
Math
Not comparable- Hy3 Preview
- Not measured
- Kimi K3
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
Multilingual
Not comparable- Hy3 Preview
- Not measured
- Kimi K3
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
Multimodal
Not comparable- Hy3 Preview
- Not measured
- Kimi K3
- 78.5
- Weighted basis
- 0 vs 3 rows
- Reading
- Not comparable
Instruction following
Not comparable- Hy3 Preview
- Not measured
- Kimi K3
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
| Category | Hy3 Preview | Kimi K3 | Weighted basis | Reading |
|---|---|---|---|---|
| Agentic | 54.4 | 89.5 | Directional only1 vs 2 rows | Directional only |
| Knowledge | 87.2 | 61.0 | Directional only1 vs 2 rows | Directional only |
| Coding | 74.4 | Not measured | Not comparable1 vs 0 rows | Not comparable |
| Reasoning | Not measured | Not measured | Not comparable0 vs 0 rows | Not comparable |
| Math | Not measured | Not measured | Not comparable0 vs 0 rows | Not comparable |
| Multilingual | Not measured | Not measured | Not comparable0 vs 0 rows | Not comparable |
| Multimodal | Not measured | 78.5 | Not comparable0 vs 3 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
GPQA
Knowledge
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
- Hy3 Preview
- Self-hosted; infrastructure cost varies
- Fits in one request
- Kimi K3
- $0.0105
- Fits in one request
Hy3 Preview has no comparable published API token rate.
Repository review
50K fresh input + 3K output tokens
- Hy3 Preview
- Self-hosted; infrastructure cost varies
- Fits in one request
- Kimi K3
- $0.195
- Fits in one request
Hy3 Preview has no comparable published API token rate.
Cache-heavy agent loop
200K cached + 20K fresh input + 10K output tokens
- Hy3 Preview
- Self-hosted; infrastructure cost varies
- Fits in one request
- Cached-input rate unavailable
- Kimi K3
- $0.27
- Fits in one request
Hy3 Preview has no comparable published API token rate.
Specification differences
Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.
SpecificationHy3 PreviewKimi K3
Context window
Maximum documented context; output-token limits may be lower.
Hy3 Preview
256K
Kimi K3
1.05M
API model ID
Hy3 Preview
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.
Hy3 Preview
No comparable hosted API rate
Kimi K3
$0.3 per 1M cached input tokens
Documented inputs
Hy3 Preview
Not sourced
Kimi K3
Not sourced
Documented outputs
Hy3 Preview
Not sourced
Kimi K3
Not sourced
Provider availability
Hy3 Preview
Not sourced
Kimi K3
Not sourced
Reasoning profile
Hy3 Preview
Reasoning
Kimi K3
Reasoning
Weight access
Hy3 Preview
Open Weight
Kimi K3
Pending
License
Hy3 Preview
Open Weight
Kimi K3
Pending
Release date
Hy3 Preview
2026-04-23
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 43.9, and the 90% score intervals do not overlap.
- Workload cost
- A complete comparable API-rate estimate is not available for both models.
- 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 evidence43 rows
Agentic
Terminal-Bench 2.0
Hy3 Preview54.4%SourceKimi K388.3%SourceKimi K3 leads this result
Gert Labs
Hy3 Preview36.91%SourceKimi K3—Not directly comparable
BrowseComp
Hy3 Preview—Kimi K391.2%SourceNot directly comparable
DeepSearchQA
Hy3 Preview—Kimi K395.0%SourceNot directly comparable
Toolathlon-Verified
Hy3 Preview—Kimi K373.2%SourceNot directly comparable
MCP Atlas
Hy3 Preview—Kimi K384.2%SourceNot directly comparable
AutomationBench
Hy3 Preview—Kimi K330.8%SourceNot directly comparable
JobBench
Hy3 Preview—Kimi K352.9%SourceNot directly comparable
APEX-Agents
Hy3 Preview—Kimi K337.6%SourceNot directly comparable
SpreadsheetBench 2
Hy3 Preview—Kimi K334.8%SourceNot directly comparable
DECK-Bench
Hy3 Preview—Kimi K373.5%SourceNot directly comparable
Coding
SWE-bench Verified
Hy3 Preview74.4%SourceKimi K3—Not directly comparable
Terminal-Bench 2.0
Hy3 Preview54.4%SourceKimi K3—Not directly comparable
deepSwe
Hy3 Preview—Kimi K367.5%SourceNot directly comparable
cursorBench32
Hy3 Preview—Kimi K360.8%SourceNot directly comparable
FrontierSWE
Hy3 Preview—Kimi K381.2%SourceNot directly comparable
ProgramBench
Hy3 Preview—Kimi K377.8%SourceNot directly comparable
Kimi Code Bench v2
Hy3 Preview—Kimi K372.9%SourceNot directly comparable
sweMarathon
Hy3 Preview—Kimi K342%SourceNot directly comparable
PostTrain Bench
Hy3 Preview—Kimi K336.6%SourceNot directly comparable
MLS-Bench Lite
Hy3 Preview—Kimi K348.3%SourceNot directly comparable
VulcanBench v3
Hy3 Preview—Kimi K373.9%SourceNot directly comparable
APEX-SWE
Hy3 Preview—Kimi K348.0%SourceNot directly comparable
EEBench
Hy3 Preview—Kimi K338.3%SourceNot directly comparable
InferenceEval
Hy3 Preview—Kimi K339.8%SourceNot directly comparable
KernelBench Internal
Hy3 Preview—Kimi K368.5%SourceNot directly comparable
Knowledge
Multimodal
OfficeQA Pro
Hy3 Preview—Kimi K363.3%SourceNot directly comparable
MMMU-Pro
Hy3 Preview—Kimi K381.6%SourceNot directly comparable
MMMU-Pro w/ Python
Hy3 Preview—Kimi K383.4%SourceNot directly comparable
CharXiv w/o tools
Hy3 Preview—Kimi K384.8%SourceNot directly comparable
CharXiv
Hy3 Preview—Kimi K391.3%SourceNot directly comparable
MathVision
Hy3 Preview—Kimi K394.3%SourceNot directly comparable
MathVision w/ Python
Hy3 Preview—Kimi K397.8%SourceNot directly comparable
BabyVision w/ Python
Hy3 Preview—Kimi K385.7%SourceNot directly comparable
ZeroBench
Hy3 Preview—Kimi K323.0%SourceNot directly comparable
ZeroBench w/ Python
Hy3 Preview—Kimi K341.0%SourceNot directly comparable
WorldVQA ForceAnswer
Hy3 Preview—Kimi K351.0%SourceNot directly comparable
OmniDocBench
Hy3 Preview—Kimi K391.1%SourceNot directly comparable
PerceptionBench
Hy3 Preview—Kimi K358.5%SourceNot directly comparable
Frequently asked questions
Which is better, Hy3 Preview or Kimi K3?
Kimi K3 has the higher public score, 80.53 versus 43.9, 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, Hy3 Preview 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, Hy3 Preview 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, Hy3 Preview or Kimi K3?
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, Hy3 Preview or Kimi K3?
Kimi K3 has the larger documented context window: 1.05M, compared with 256K.
Related comparisons
Last updated August 19, 2026
Watch Hy3 Preview 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.