Capability
Unranked
field median 57.5
Not eligible for a public rank
DeepSeek V4 Flash 0731
Released Jul 31, 2026 — see all recent releases
Decision reading
DeepSeek V4 Flash 0731 is tracked, but not publicly ranked yet. The profile exposes 33 sourced benchmark rows and leaves unsupported fields blank until a published record exists.
Data as of August 13, 2026 · How the score is built
Strongest published evidence
Knowledge ranks #41. Particularly effective for knowledge-intensive tasks like research, analysis, and factual Q&A.
Validate before choosing
33 published rows leave some tracked benchmark slots empty. Independent runtime speed has not been measured.
Decision snapshot
Each value carries a field reference instead of floating alone. Markers compare this model with the current ranked and priced catalog; they are not absolute quality thresholds.
Price
$0.14input / $0.28 output
input median $1
cached $0.003 · blended $0.21
Speed
Not measured
field median 93 tok/s
Time to first token not measured
Context
1Mtokens
field median 200,000
Maximum output length is tracked separately
How much of this is verified
Coverage is split by category so a strong number never hides a thin evidence base. Verified means the row is tied to a published source; provisional rows remain visible but separate.
- Agentic10/10 verified
- Coding11/11 verified
- Reasoning2/2 verified
- Knowledge6/6 verified
- Math4/4 verified
- MultilingualNot measured
- MultimodalNot measured
- Inst. FollowingNot measured
Verified sourceProvisionalNot measured
Spec sheet
Each documented value carries its source. Missing fields stay visible as not sourced or not published, rather than disappearing from the page.
- API model ID
- deepseek-v4-flashDeepSeek V4 Flash 0731 update
- Context window
- 1MDeepSeek models and pricing
- Maximum output
- Not sourced yet
- Knowledge cutoff
- Not sourced yet
- Input modalities
- textDeepSeek GitHub Copilot integration
- Output modalities
- textDeepSeek GitHub Copilot integration
- Parameters
- Not disclosed by the provider
- Availability
- DeepSeek APIDeepSeek V4 Flash 0731 update
- Cloud regions
- Not tracked yet
- Lifecycle
- activeDeepSeek V4 Flash 0731 update
- API capabilities
- Tool calling, structured outputs, and batch support are not tracked yet
- Prompt caching
- Published at $0.003 per million cached input tokens
- Self-host
- Weights are not published
- Rate limits
- Not tracked yet
Category score record
Scores and ranks appear only where published evidence can be displayed. The table keeps the score, weight, cohort, and evidence state together.
| Category | Score | Rank | Percentile | Weight | Benchmarks | Evidence |
|---|---|---|---|---|---|---|
| AgenticRank Not rankedWeight 22%10 benchmarksVerified | 49.8 | Not ranked | Not available | 22% | 10 benchmarks | Verified |
| CodingRank Not rankedWeight 20%11 benchmarksVerified | 46.6 | Not ranked | Not available | 20% | 11 benchmarks | Verified |
| ReasoningWeight 17%2 benchmarksVerified | Score pending | Not ranked | Not available | 17% | 2 benchmarks | Verified |
| KnowledgeRank #41 of 55Percentile 26thWeight 12%6 benchmarksVerified | 60.1 | #41 of 55 | 26th | 12% | 6 benchmarks | Verified |
| MathRank Not rankedWeight 5%4 benchmarksVerified | 81.4 | Not ranked | Not available | 5% | 4 benchmarks | Verified |
| MultilingualWeight 7%0 benchmarksNot measured | Not measured | Not ranked | Not available | 7% | 0 benchmarks | Not measured |
| MultimodalWeight 12%0 benchmarksNot measured | Not measured | Not ranked | Not available | 12% | 0 benchmarks | Not measured |
| Inst. FollowingWeight 5%0 benchmarksNot measured | Not measured | Not ranked | Not available | 5% | 0 benchmarks | Not measured |
Benchmark ledger
Coding opens by default. The marker compares each value with the best source-verified result in the catalog; provisional leaders do not set the reference. Expand the remaining categories for every published row.
Coding11 rows
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| SWE-bench VerifiedSoftware Engineering Benchmark Verified | Score79% | Versus best verified row Best verified: Claude Opus 5 · 96% | Gap17 behind | WeightWeighted 16% | Provider exact |
| SWE-bench Pro | Score52.6% | Versus best verified row Best verified: Claude Mythos 5 · 80.3% | Gap27.7 behind | WeightWeighted 10% | Provider exact |
| LiveCodeBench Pass@1-COTLiveCodeBench Pass@1 with Chain-of-Thought | Score91.6% | Versus best verified row Best verified: DeepSeek V4 Pro 0813 · 93.5% | Gap1.9 behind | WeightDisplay only | Provider exact |
| CodeforcesCodeforces Rating | Score3052.0 | Versus best verified row Best verified: DeepSeek V4 Pro 0813 · 3206.0 | Gap154 behind | WeightDisplay only | Provider exact |
| SWE Multilingual | Score73.3% | Versus best verified row Best verified: Claude Opus 5 · 89.5% | Gap16.2 behind | WeightDisplay only | Provider exact |
| Terminal-Bench 2.0 | Score56.9% | Versus best verified row Best verified: GPT-5.6 Sol · 91.9% | Gap35 behind | WeightDisplay only | Provider exact |
| Terminal-Bench 2.1Terminal-Bench 2.1 (provider run) | Score82.7% | Versus best verified row Best verified: DeepSeek V4 Pro 0813 · 87.9% | Gap5.2 behind | WeightDisplay only | Provider exact |
| NL2Repo | Score54.2% | Versus best verified row Best verified: DeepSeek V4 Pro 0813 · 61.5% | Gap7.3 behind | WeightDisplay only | Provider exact |
| deepSwe | Score54.4% | Versus best verified row Best verified: GPT-5.6 Sol · 72.7% | Gap18.3 behind | WeightDisplay only | Provider exact |
| DSBench-FullStackDeepSeek DSBench FullStack | Score68.7% | Versus best verified row Best verified: DeepSeek V4 Pro 0813 · 71.1% | Gap2.4 behind | WeightDisplay only | Provider exact |
| DSBench-HardDeepSeek DSBench Hard | Score59.6% | Versus best verified row Best verified: DeepSeek V4 Pro 0813 · 67.2% | Gap7.6 behind | WeightDisplay only | Provider exact |
Agentic10 rows
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| Terminal-Bench 2.0 | Score56.9% | Versus best verified row Best verified: GPT-5.6 Sol · 91.9% | Gap35 behind | WeightWeighted 38% | Provider exact |
| BrowseComp | Score73.2% | Versus best verified row Best verified: GPT-5.6 Sol · 92.2% | Gap19 behind | WeightWeighted 28% | Provider exact |
| HLE w/ toolsHumanity's Last Exam with tools | Score45.1% | Versus best verified row Best verified: Claude Opus 5 · 64.7% | Gap19.6 behind | WeightDisplay only | Provider exact |
| MCP Atlas | Score69% | Versus best verified row Best verified: Muse Spark 1.1 · 88.1% | Gap19.1 behind | WeightDisplay only | Provider exact |
| Toolathlon | Score47.8% | Versus best verified row Best verified: Muse Spark 1.1 · 75.6% | Gap27.8 behind | WeightDisplay only | Provider exact |
| Terminal-Bench 2.1Terminal-Bench 2.1 (provider run) | Score82.7% | Versus best verified row Best verified: DeepSeek V4 Pro 0813 · 87.9% | Gap5.2 behind | WeightDisplay only | Provider exact |
| CyberGym | Score76.7% | Versus best verified row Best verified: Fugu Cyber · 86.9% | Gap10.2 behind | WeightDisplay only | Provider exact |
| Toolathlon-Verified | Score70.3% | Versus best verified row Best verified: Claude Opus 5 · 80.6% | Gap10.3 behind | WeightDisplay only | Provider exact |
| Agents' Last Exam | Score25.2% | Versus best verified row Best verified: Qwen3.8 Max · 52.4% | Gap27.2 behind | WeightDisplay only | Provider exact |
| AutomationBench | Score25.1% | Versus best verified row Best verified: DeepSeek V4 Pro 0813 · 31.8% | Gap6.7 behind | WeightDisplay only | Provider exact |
Reasoning2 rows
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| MRCR 1M | Score78.7% | Versus best verified row Best verified: DeepSeek V4 Pro 0813 · 83.5% | Gap4.8 behind | WeightDisplay only | Provider exact |
| CorpusQA 1M | Score60.5% | Versus best verified row Best verified: DeepSeek V4 Pro 0813 · 62.0% | Gap1.5 behind | WeightDisplay only | Provider exact |
Knowledge6 rows
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| HLEHumanity's Last Exam | Score34.8% | Versus best verified row Best verified: Claude Opus 5 · 64.7% | Gap29.9 behind | WeightWeighted 45% | Provider exact |
| MMLU-ProMassive Multitask Language Understanding Professional | Score86.2% | Versus best verified row Best verified: Qwen3.7 Max · 89.6% | Gap3.4 behind | WeightWeighted 30% | Provider exact |
| SimpleQAMeasuring Short-Form Factuality in Large Language Models | Score34.1% | Versus best verified row Best verified: DeepSeek V4 Pro 0813 · 57.9% | Gap23.8 behind | WeightWeighted 11% | Provider exact |
| GPQAGraduate-Level Google-Proof Q&A | Score88.1% | Versus best verified row Best verified: Sakana Fugu-Ultra · 95.5% | Gap7.4 behind | WeightWeighted 7% | Provider exact |
| Chinese-SimpleQA | Score78.9% | Versus best verified row Best verified: DeepSeek V4 Pro 0813 · 84.4% | Gap5.5 behind | WeightDisplay only | Provider exact |
| GPQA-DGPQA Diamond | Score88.1% | Versus best verified row Best verified: Sakana Fugu-Ultra · 95.5% | Gap7.4 behind | WeightDisplay only | Provider exact |
Math4 rows
| Benchmark | Score | Versus best verified row | Gap | Weight | Evidence |
|---|---|---|---|---|---|
| HMMT Feb 2026Harvard-MIT Mathematics Tournament February 2026 | Score94.8% | Versus best verified row Best verified: Qwen3.7 Max · 97.1% | Gap2.3 behind | WeightWeighted 25% | Provider exact |
| IMOAnswerBench | Score88.4% | Versus best verified row Best verified: Qwen3.7 Max · 90.0% | Gap1.6 behind | WeightDisplay only | Provider exact |
| Apex | Score33.0% | Versus best verified row Best verified: Qwen3.7 Max · 44.5% | Gap11.5 behind | WeightDisplay only | Provider exact |
| Apex Shortlist | Score85.7% | Versus best verified row Best verified: DeepSeek V4 Pro 0813 · 90.2% | Gap4.5 behind | WeightDisplay only | Provider exact |
Lineage
The sequence follows explicit supersedes links. Scores and prices remain blank when the corresponding public row or first-party rate is unavailable.
Jul 31, 2026 · you are here
DeepSeek V4 Flash 0731Not publicly ranked · $0.14 / $0.28
flash-reasoning · 0731
How to read this profile
The visual layer above carries the decisions. These notes preserve the model, ranking, coverage, and family context behind the numbers.
We track DeepSeek V4 Flash 0731, but the public leaderboard excludes this profile until enough non-generated benchmark coverage is available. Only published rows appear above.
DeepSeek V4 Flash 0731 is a proprietary model with a 1M context window. It uses an explicit reasoning mode, which can improve complex problem solving while adding latency and token use.
The existing deepseek-v4-flash API ID now serves DeepSeek-V4-Flash-0731 in public beta. DeepSeek says it has the same architecture and size as Preview with new post-training. Checked again August 1: DeepSeek-V4-Flash-0731 weights are public on Hugging Face under the MIT License, so this row is Open Weight. The Hugging Face card lists 304B parameters because 0731 ships with the DSpark speculative-decoding module attached; the April Preview weights remain a separate checkpoint.
Its explicit predecessor is DeepSeek V4 Flash 0731. 33 of 414 tracked benchmark slots currently have displayable evidence. Missing categories stay blank.
Its strongest eligible category is Knowledge at #41. particularly effective for knowledge-intensive tasks like research, analysis, and factual Q&A.
Radar
Full release historyDeepSeek V4 Flash 0731 release history
DeepSeek V4 Flash 0731 released
DeepSeek · Model release
Source confirmed
Frequently asked questions
How does DeepSeek V4 Flash 0731 perform overall in AI benchmarks?
DeepSeek V4 Flash 0731 has 33 source-displayable benchmark rows, but it does not qualify for a public overall rank. The available rows remain visible by category without being converted into a site-wide score. Missing evidence stays blank instead of being estimated from an earlier model.
Is DeepSeek V4 Flash 0731 good for knowledge and understanding?
DeepSeek V4 Flash 0731 ranks #41 out of 55 eligible models for knowledge and understanding, with a public category score of 60.1/100. Higher-ranked alternatives are available for workloads where this category decides the choice. Check the underlying rows before treating the aggregate as a workload guarantee.
Is DeepSeek V4 Flash 0731 good for coding and programming?
DeepSeek V4 Flash 0731 has source-displayable benchmark coverage for coding and programming, but the public category table does not assign it a rank there. The individual rows remain available for inspection. A missing category position means the evidence threshold was not met; it does not convert the model's unmeasured work into a zero.
Is DeepSeek V4 Flash 0731 good for mathematics?
DeepSeek V4 Flash 0731 has source-displayable benchmark coverage for mathematics, but the public category table does not assign it a rank there. The individual rows remain available for inspection. A missing category position means the evidence threshold was not met; it does not convert the model's unmeasured work into a zero.
Is DeepSeek V4 Flash 0731 good for reasoning and logic?
DeepSeek V4 Flash 0731 has source-displayable benchmark coverage for reasoning and logic, but the public category table does not assign it a rank there. The individual rows remain available for inspection. A missing category position means the evidence threshold was not met; it does not convert the model's unmeasured work into a zero.
Is DeepSeek V4 Flash 0731 good for agentic tool use and computer tasks?
DeepSeek V4 Flash 0731 has source-displayable benchmark coverage for agentic tool use and computer tasks, but the public category table does not assign it a rank there. The individual rows remain available for inspection. A missing category position means the evidence threshold was not met; it does not convert the model's unmeasured work into a zero.
Does DeepSeek V4 Flash 0731 have full benchmark coverage on BenchLM?
No. DeepSeek V4 Flash 0731 currently has 47 source-displayable rows across 414 tracked benchmark slots. The profile exposes published, non-generated evidence and leaves missing categories blank until an exact evaluation is available. Coverage describes how much was measured; it is not a penalty added to an individual benchmark result.
What is the context window size of DeepSeek V4 Flash 0731?
DeepSeek V4 Flash 0731 has a documented context window of 1M. That figure is the maximum combined prompt and retained-conversation space reported for this exact model; it is not the maximum output length. The profile keeps output limits separate because providers often publish those limits independently.
Last updated August 13, 2026. Runtime fields remain blank until a sourced snapshot exists.
Watch DeepSeek V4 Flash 0731 in the weekly brief
Get one weekly email when material rank, price, availability, or benchmark evidence changes are worth revisiting.
Read a sample issueJoin 2,000+ readers.