Coding
Directional only- DeepSeek V4 Flash 0731
- 68.8
- Hy4 preview
- 65.7
- Weighted basis
- 2 vs 1 rows
- Reading
- Directional only
DeepSeek V4 Flash 0731 vs Hy4 preview
Updated August 28, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.
Decision reading
At least one model is not scored in the current public ranking lane, so the page does not name an overall quality winner.
16 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.
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
The category averages use different weighted benchmark sets, so they are directional rather than like-for-like.
Confidence: limited
Agentic work
Tool use, computer use, and multi-step task completion
Not enough matched evidence
No shared weighted benchmark basis supports a winner.
Confidence: limited
Long documents
Prompts that approach the documented context limit
No clear pick
The documented context windows are equal.
Confidence: documented
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.
19 DeepSeek V4 Flash 0731 only
16 Shared
12 Hy4 preview only
- Shared results
- 16
- DeepSeek V4 Flash 0731 only
- 19
- Hy4 preview only
- 12
- 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- DeepSeek V4 Flash 0731
- 55.3
- Hy4 preview
- 60.4
- Weighted basis
- 4 vs 2 rows
- Reading
- Directional only
Agentic
Not comparable- DeepSeek V4 Flash 0731
- 63.8
- Hy4 preview
- Not measured
- Weighted basis
- 2 vs 0 rows
- Reading
- Not comparable
Reasoning
Not comparable- DeepSeek V4 Flash 0731
- Not measured
- Hy4 preview
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
Math
Not comparable- DeepSeek V4 Flash 0731
- 94.8
- Hy4 preview
- Not measured
- Weighted basis
- 1 vs 0 rows
- Reading
- Not comparable
Multilingual
Not comparable- DeepSeek V4 Flash 0731
- Not measured
- Hy4 preview
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
Multimodal
Not comparable- DeepSeek V4 Flash 0731
- Not measured
- Hy4 preview
- 66.2
- Weighted basis
- 0 vs 1 rows
- Reading
- Not comparable
Instruction following
Not comparable- DeepSeek V4 Flash 0731
- Not measured
- Hy4 preview
- Not measured
- Weighted basis
- 0 vs 0 rows
- Reading
- Not comparable
| Category | DeepSeek V4 Flash 0731 | Hy4 preview | Weighted basis | Reading |
|---|---|---|---|---|
| Coding | 68.8 | 65.7 | Directional only2 vs 1 rows | Directional only |
| Knowledge | 55.3 | 60.4 | Directional only4 vs 2 rows | Directional only |
| Agentic | 63.8 | Not measured | Not comparable2 vs 0 rows | Not comparable |
| Reasoning | Not measured | Not measured | Not comparable0 vs 0 rows | Not comparable |
| Math | 94.8 | Not measured | Not comparable1 vs 0 rows | Not comparable |
| Multilingual | Not measured | Not measured | Not comparable0 vs 0 rows | Not comparable |
| Multimodal | Not measured | 66.2 | Not comparable0 vs 1 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.
HLE
Knowledge
DeepSeek V4 Flash 0731: 34.8%Hy4 preview: 55.4%Normalized gap 20.6DeepSeek V4 Flash 0731 source Hy4 preview sourceSWE-bench Pro
Coding
DeepSeek V4 Flash 0731: 52.6%Hy4 preview: 65.7%Normalized gap 13.1DeepSeek V4 Flash 0731 source Hy4 preview sourceGPQA
Knowledge
DeepSeek V4 Flash 0731: 88.1%Hy4 preview: 92.3%Normalized gap 4.2DeepSeek V4 Flash 0731 source Hy4 preview source
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
- DeepSeek V4 Flash 0731
- $0.00028
- Fits in one request
- Hy4 preview
- Self-hosted; infrastructure cost varies
- Fits in one request
Hy4 preview has no comparable published API token rate.
Repository review
50K fresh input + 3K output tokens
- DeepSeek V4 Flash 0731
- $0.00784
- Fits in one request
- Hy4 preview
- Self-hosted; infrastructure cost varies
- Fits in one request
Hy4 preview has no comparable published API token rate.
Cache-heavy agent loop
200K cached + 20K fresh input + 10K output tokens
- DeepSeek V4 Flash 0731
- $0.00616
- Fits in one request
- Hy4 preview
- Self-hosted; infrastructure cost varies
- Fits in one request
- Cached-input rate unavailable
Hy4 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.
SpecificationDeepSeek V4 Flash 0731Hy4 preview
Context window
Maximum documented context; output-token limits may be lower.
DeepSeek V4 Flash 0731
Hy4 preview
API model ID
DeepSeek V4 Flash 0731
deepseek-v4-flash
DeepSeek V4 Flash 0731 updateHy4 preview
tencent/Hy4-preview
Tencent Hy4 preview model cardCached-input rate
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
DeepSeek V4 Flash 0731
$0.0028 per 1M cached input tokens
Hy4 preview
No comparable hosted API rate
Tencent Hy4 preview model cardDocumented inputs
DeepSeek V4 Flash 0731
Hy4 preview
Not sourced
Documented outputs
DeepSeek V4 Flash 0731
Hy4 preview
Not sourced
Provider availability
DeepSeek V4 Flash 0731
Public Beta · DeepSeek API
DeepSeek V4 Flash 0731 updateHy4 preview
Not sourced
Reasoning profile
DeepSeek V4 Flash 0731
Reasoning
Hy4 preview
Reasoning
Weight access
DeepSeek V4 Flash 0731
Proprietary
Hy4 preview
Open Weight
License
DeepSeek V4 Flash 0731
Proprietary
Hy4 preview
Open Weight
Release date
DeepSeek V4 Flash 0731
2026-07-31
Hy4 preview
2026-08-28
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
- At least one model is not scored in the current public ranking lane, so the page does not name an overall quality winner.
- Workload cost
- A complete comparable API-rate estimate is not available for both models.
- Context tradeoff
- Both models list 1M.
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 evidence47 rows
Agentic
Terminal-Bench 2.0
DeepSeek V4 Flash 073156.9%SourceHy4 preview—Not directly comparable
BrowseComp
DeepSeek V4 Flash 073173.2%SourceHy4 preview—Not directly comparable
HLE w/ tools
DeepSeek V4 Flash 073145.1%SourceHy4 preview55.4%SourceHy4 preview leads this result
MCP Atlas
DeepSeek V4 Flash 073169%SourceHy4 preview83.7%SourceHy4 preview leads this result
Toolathlon
DeepSeek V4 Flash 073147.8%SourceHy4 preview—Not directly comparable
Terminal-Bench 2.1
DeepSeek V4 Flash 073182.7%SourceHy4 preview85.4%SourceHy4 preview leads this result
CyberGym
DeepSeek V4 Flash 073176.7%SourceHy4 preview78.4%SourceHy4 preview leads this result
Toolathlon-Verified
DeepSeek V4 Flash 073170.3%SourceHy4 preview74.1%SourceHy4 preview leads this result
Agents' Last Exam
DeepSeek V4 Flash 073125.2%SourceHy4 preview22.8%SourceDeepSeek V4 Flash 0731 leads this result
AutomationBench
DeepSeek V4 Flash 073125.1%SourceHy4 preview32.1%SourceHy4 preview leads this result
WideResearch
DeepSeek V4 Flash 0731—Hy4 preview83.9%SourceNot directly comparable
DRACO
DeepSeek V4 Flash 0731—Hy4 preview77.2%SourceNot directly comparable
APEX-Agents
DeepSeek V4 Flash 0731—Hy4 preview37.1%SourceNot directly comparable
skillsBench
DeepSeek V4 Flash 0731—Hy4 preview62.9%SourceNot directly comparable
JobBench
DeepSeek V4 Flash 0731—Hy4 preview61.7%SourceNot directly comparable
BankerToolBench
DeepSeek V4 Flash 0731—Hy4 preview78.6%SourceNot directly comparable
Coding
LiveCodeBench Pass@1-COT
DeepSeek V4 Flash 073191.6%SourceHy4 preview—Not directly comparable
Codeforces
DeepSeek V4 Flash 07313052.0SourceHy4 preview—Not directly comparable
SWE-bench Verified
DeepSeek V4 Flash 073179%SourceHy4 preview—Not directly comparable
SWE-bench Pro
DeepSeek V4 Flash 073152.6%SourceHy4 preview65.7%SourceHy4 preview leads this result
SWE Multilingual
DeepSeek V4 Flash 073173.3%SourceHy4 preview82.9%SourceHy4 preview leads this result
Terminal-Bench 2.0
DeepSeek V4 Flash 073156.9%SourceHy4 preview—Not directly comparable
Terminal-Bench 2.1
DeepSeek V4 Flash 073182.7%SourceHy4 preview85.4%SourceHy4 preview leads this result
NL2Repo
DeepSeek V4 Flash 073154.2%SourceHy4 preview58.9%SourceHy4 preview leads this result
deepSwe
DeepSeek V4 Flash 073154.4%SourceHy4 preview64.3%SourceHy4 preview leads this result
DSBench-FullStack
DeepSeek V4 Flash 073168.7%SourceHy4 preview—Not directly comparable
DSBench-Hard
DeepSeek V4 Flash 073159.6%SourceHy4 preview—Not directly comparable
VulcanBench v3
DeepSeek V4 Flash 073188.4%SourceHy4 preview—Not directly comparable
OpenHarmony Bench
DeepSeek V4 Flash 073153.8%SourceHy4 preview—Not directly comparable
ProgramBench
DeepSeek V4 Flash 0731—Hy4 preview17.5%SourceNot directly comparable
PostTrain Bench
DeepSeek V4 Flash 0731—Hy4 preview35.6%SourceNot directly comparable
sweMarathon
DeepSeek V4 Flash 0731—Hy4 preview31.9%SourceNot directly comparable
Reasoning
Knowledge
MMLU-Pro
DeepSeek V4 Flash 073186.2%SourceHy4 preview—Not directly comparable
SimpleQA
DeepSeek V4 Flash 073134.1%SourceHy4 preview—Not directly comparable
Chinese-SimpleQA
DeepSeek V4 Flash 073178.9%SourceHy4 preview—Not directly comparable
GPQA
DeepSeek V4 Flash 073188.1%SourceHy4 preview92.3%SourceHy4 preview leads this result
GPQA-D
DeepSeek V4 Flash 073188.1%SourceHy4 preview92.3%SourceHy4 preview leads this result
HLE
DeepSeek V4 Flash 073134.8%SourceHy4 preview55.4%SourceHy4 preview leads this result
HLE w/o tools
DeepSeek V4 Flash 0731—Hy4 preview43.4%SourceNot directly comparable
Math
HMMT Feb 2026
DeepSeek V4 Flash 073194.8%SourceHy4 preview—Not directly comparable
IMOAnswerBench
DeepSeek V4 Flash 073188.4%SourceHy4 preview—Not directly comparable
Apex
DeepSeek V4 Flash 073133.0%SourceHy4 preview74.2%SourceHy4 preview leads this result
Apex Shortlist
DeepSeek V4 Flash 073185.7%SourceHy4 preview—Not directly comparable
Multimodal
OfficeQA Pro
DeepSeek V4 Flash 0731—Hy4 preview66.2%SourceNot directly comparable
Frequently asked questions
Which is better, DeepSeek V4 Flash 0731 or Hy4 preview?
At least one model is not scored in the current public ranking lane, so the page does not name an overall quality winner. The page therefore keeps the decision tied to the specific documented workload.
Which is better for coding, DeepSeek V4 Flash 0731 or Hy4 preview?
The current coding 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 is better for agentic tasks, DeepSeek V4 Flash 0731 or Hy4 preview?
The published evidence does not provide a shared weighted agentic tasks basis for both models, so BenchLM does not name a agentic tasks winner.
Which costs less, DeepSeek V4 Flash 0731 or Hy4 preview?
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, DeepSeek V4 Flash 0731 or Hy4 preview?
Both models list the same context window, 1M.
Related comparisons
Last updated August 28, 2026
Watch DeepSeek V4 Flash 0731 vs Hy4 preview
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 2,000+ readers.