Agentic
Not comparable- Hy ASR 3.0 preview
- Not ranked
- Kimi K3
- 71.9
- Supported · #4/152
- Basis
- BenchAlign lane · 0 vs 11 public rows
- Reading
- Not comparable
Hy ASR 3.0 preview vs Kimi K3
Updated September 10, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.
Decision reading
The public evidence has no benchmark result shared by both models, so it does not support a quality verdict. Use the documented cost, context, and runtime rows instead.
0 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.
Share or export
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.
No workload recommendation clears the current evidence threshold.
Use the matched evidence, workload costs, and sourced specifications below instead of treating a point score as a universal answer.
Show secondary and unsupported calls
Coding work
Code generation, repair, and software-engineering tasks
Not enough matched evidence
Hy ASR 3.0 preview is not ranked on the public lane for coding, so no winner is named for coding.
Confidence: limited
Agentic work
Tool use, computer use, and multi-step task completion
Not enough matched evidence
Hy ASR 3.0 preview is not ranked on the public lane for agentic, so no winner is named for agentic.
Confidence: limited
Long documents
Prompts that approach the documented context limit
Not enough matched evidence
A complete context comparison is not sourced.
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.
43 Kimi K3 only
- Shared results
- 0
- Hy ASR 3.0 preview only
- 0
- Kimi K3 only
- 43
- Like-for-like categories
- 0 / 8
Category results, on a stated basis
Each row shows the public-lane category score for both models: the BenchAlign 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
Not comparable- Hy ASR 3.0 preview
- Not ranked
- Kimi K3
- 68.0
- Supported · #6/151
- Basis
- BenchAlign lane · 0 vs 13 public rows
- Reading
- Not comparable
Reasoning
Not comparable- Hy ASR 3.0 preview
- Not ranked
- Kimi K3
- 78.5
- #3/20
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Knowledge
Not comparable- Hy ASR 3.0 preview
- Not ranked
- Kimi K3
- 72.0
- Supported · #8/183
- Basis
- BenchAlign lane · 0 vs 6 public rows
- Reading
- Not comparable
Math
Not comparable- Hy ASR 3.0 preview
- Not ranked
- Kimi K3
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Multilingual
Not comparable- Hy ASR 3.0 preview
- Not ranked
- Kimi K3
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
Multimodal
Not comparable- Hy ASR 3.0 preview
- Not ranked
- Kimi K3
- 89.5
- #1/48
- Basis
- Provisional lane · 0 vs 3 weighted rows
- Reading
- Not comparable
Instruction following
Not comparable- Hy ASR 3.0 preview
- Not ranked
- Kimi K3
- Not ranked
- Basis
- Provisional lane · 0 vs 0 weighted rows
- Reading
- Not comparable
| Category | Hy ASR 3.0 preview | Kimi K3 | Basis | Reading |
|---|---|---|---|---|
| Agentic | Not ranked | 71.9Supported · #4/152 | Not comparableBenchAlign lane · 0 vs 11 public rows | Not comparable |
| Coding | Not ranked | 68.0Supported · #6/151 | Not comparableBenchAlign lane · 0 vs 13 public rows | Not comparable |
| Reasoning | Not ranked | 78.5#3/20 | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Knowledge | Not ranked | 72.0Supported · #8/183 | Not comparableBenchAlign lane · 0 vs 6 public rows | Not comparable |
| Math | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Multilingual | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Multimodal | Not ranked | 89.5#1/48 | Not comparableProvisional lane · 0 vs 3 weighted rows | Not comparable |
| Instruction following | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
Ranks count the models scored in each category’s lane, so the agentic, coding, and knowledge denominators (BenchAlign) differ from the provisional-lane categories. Unranked scores sit on the lane’s scale but never name a winner. Like-for-like rows within 0.5 points read as a practical tie.
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.
A shared-evidence shape is not available.
BenchLM does not draw a radar or infer missing axes when the matched evidence is too sparse.
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
- Hy ASR 3.0 preview
- API rate not published
- Fit state unavailable
- Kimi K3
- $0.0105
- Fits in one request
Hy ASR 3.0 preview has no comparable published API token rate.
Repository review
50K fresh input + 3K output tokens
- Hy ASR 3.0 preview
- API rate not published
- Fit state unavailable
- Kimi K3
- $0.195
- Fits in one request
Hy ASR 3.0 preview has no comparable published API token rate.
Cache-heavy agent loop
200K cached + 20K fresh input + 10K output tokens
- Hy ASR 3.0 preview
- API rate not published
- Fit state unavailable
- Cached-input rate unavailable
- Kimi K3
- $0.27
- Fits in one request
Hy ASR 3.0 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.
SpecificationHy ASR 3.0 previewKimi K3
Context window
Maximum documented context; output-token limits may be lower.
Hy ASR 3.0 preview
N/A
Kimi K3
1.05M
API model ID
Hy ASR 3.0 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.
Hy ASR 3.0 preview
No comparable hosted API rate
Tencent model documentationKimi K3
$0.3 per 1M cached input tokens
Documented inputs
Hy ASR 3.0 preview
Not sourced
Kimi K3
Not sourced
Documented outputs
Hy ASR 3.0 preview
Not sourced
Kimi K3
Not sourced
Provider availability
Hy ASR 3.0 preview
Not sourced
Kimi K3
Not sourced
Reasoning profile
Hy ASR 3.0 preview
Non-Reasoning
Kimi K3
Reasoning
Weight access
Hy ASR 3.0 preview
Proprietary
Kimi K3
Pending
License
Hy ASR 3.0 preview
Proprietary
Kimi K3
Pending
Release date
Hy ASR 3.0 preview
2026-08-04
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
- The public evidence has no benchmark result shared by both models, so it does not support a quality verdict.
- Workload cost
- A complete comparable API-rate estimate is not available for both models.
- Context tradeoff
- A complete documented context comparison is not available.
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
Hy ASR 3.0 preview—Kimi K388.3%SourceNot directly comparable
BrowseComp
Hy ASR 3.0 preview—Kimi K391.2%SourceNot directly comparable
DeepSearchQA
Hy ASR 3.0 preview—Kimi K395.0%SourceNot directly comparable
Toolathlon-Verified
Hy ASR 3.0 preview—Kimi K373.2%SourceNot directly comparable
MCP Atlas
Hy ASR 3.0 preview—Kimi K384.2%SourceNot directly comparable
AutomationBench
Hy ASR 3.0 preview—Kimi K330.8%SourceNot directly comparable
JobBench
Hy ASR 3.0 preview—Kimi K352.9%SourceNot directly comparable
APEX-Agents
Hy ASR 3.0 preview—Kimi K337.6%SourceNot directly comparable
SpreadsheetBench 2
Hy ASR 3.0 preview—Kimi K334.8%SourceNot directly comparable
DECK-Bench
Hy ASR 3.0 preview—Kimi K373.5%SourceNot directly comparable
Terminal-Bench 2.1 (Vals)
Hy ASR 3.0 preview—Kimi K380.9%SourceNot directly comparable
Coding
DeepSWE
Hy ASR 3.0 preview—Kimi K367.5%SourceNot directly comparable
cursorBench32
Hy ASR 3.0 preview—Kimi K360.8%SourceNot directly comparable
FrontierSWE
Hy ASR 3.0 preview—Kimi K381.2%SourceNot directly comparable
ProgramBench
Hy ASR 3.0 preview—Kimi K377.8%SourceNot directly comparable
Kimi Code Bench v2
Hy ASR 3.0 preview—Kimi K372.9%SourceNot directly comparable
sweMarathon
Hy ASR 3.0 preview—Kimi K342%SourceNot directly comparable
PostTrain Bench
Hy ASR 3.0 preview—Kimi K336.6%SourceNot directly comparable
MLS-Bench Lite
Hy ASR 3.0 preview—Kimi K348.3%SourceNot directly comparable
VulcanBench v3
Hy ASR 3.0 preview—Kimi K373.7%SourceNot directly comparable
OpenHarmony Bench
Hy ASR 3.0 preview—Kimi K357.3%SourceNot directly comparable
FrontierSWE v2
Hy ASR 3.0 preview—Kimi K325.9%SourceNot directly comparable
LiveCodeBench (Vals)
Hy ASR 3.0 preview—Kimi K387.2%SourceNot directly comparable
SWE-bench (Vals)
Hy ASR 3.0 preview—Kimi K393.4%SourceNot directly comparable
Knowledge
GPQA
Hy ASR 3.0 preview—Kimi K393.5%SourceNot directly comparable
GPQA-D
Hy ASR 3.0 preview—Kimi K393.5%SourceNot directly comparable
HLE
Hy ASR 3.0 preview—Kimi K356%SourceNot directly comparable
HLE w/o tools
Hy ASR 3.0 preview—Kimi K343.5%SourceNot directly comparable
GPQA Diamond (Vals)
Hy ASR 3.0 preview—Kimi K392.9%SourceNot directly comparable
MMLU-Pro (Vals)
Hy ASR 3.0 preview—Kimi K388.0%SourceNot directly comparable
Multimodal
OfficeQA Pro
Hy ASR 3.0 preview—Kimi K363.3%SourceNot directly comparable
MMMU-Pro
Hy ASR 3.0 preview—Kimi K381.6%SourceNot directly comparable
MMMU-Pro w/ Python
Hy ASR 3.0 preview—Kimi K383.4%SourceNot directly comparable
CharXiv w/o tools
Hy ASR 3.0 preview—Kimi K384.8%SourceNot directly comparable
CharXiv
Hy ASR 3.0 preview—Kimi K391.3%SourceNot directly comparable
MathVision
Hy ASR 3.0 preview—Kimi K394.3%SourceNot directly comparable
MathVision w/ Python
Hy ASR 3.0 preview—Kimi K397.8%SourceNot directly comparable
BabyVision w/ Python
Hy ASR 3.0 preview—Kimi K385.7%SourceNot directly comparable
ZeroBench
Hy ASR 3.0 preview—Kimi K323.0%SourceNot directly comparable
ZeroBench w/ Python
Hy ASR 3.0 preview—Kimi K341.0%SourceNot directly comparable
WorldVQA ForceAnswer
Hy ASR 3.0 preview—Kimi K351.0%SourceNot directly comparable
OmniDocBench
Hy ASR 3.0 preview—Kimi K391.1%SourceNot directly comparable
PerceptionBench
Hy ASR 3.0 preview—Kimi K358.5%SourceNot directly comparable
Frequently asked questions
Which is better, Hy ASR 3.0 preview or Kimi K3?
The public evidence has no benchmark result shared by both models, so it does not support a quality verdict. The page therefore keeps the decision tied to the specific documented workload.
Which is better for coding, Hy ASR 3.0 preview or Kimi K3?
Hy ASR 3.0 preview is not ranked on the public lane for coding, so no winner is named for coding.
Which is better for agentic tasks, Hy ASR 3.0 preview or Kimi K3?
Hy ASR 3.0 preview is not ranked on the public lane for agentic tasks, so no winner is named for agentic tasks.
Which costs less, Hy ASR 3.0 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, Hy ASR 3.0 preview or Kimi K3?
A complete documented context-window comparison is not available.
Related comparisons
Last updated September 10, 2026
Watch Hy ASR 3.0 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.