DeepSeek V4-Flash

DeepSeek V4-Flash is the smaller of the two large language models in the DeepSeek V4 family, a 284-billion-parameter Mixture of Experts model with 13 billion active parameters and a one-million-token context window, developed by the Hangzhou lab DeepSeek.^[19][20] It first shipped on April 24, 2026 as a preview, alongside the larger V4-Pro variant of the same DeepSeek V4 release. On July 31, 2026 DeepSeek put the official build, DeepSeek-V4-Flash-0731, into public beta on its API, keeping the same architecture, the same parameter count, and the same prices as the preview while reporting much higher scores on agent benchmarks.^[1][2] The 0731 release also added native support for OpenAI's Responses API wire format and a documented configuration for OpenAI Codex.^[1][5]

Scroll sideways for more →

Background

DeepSeek released the V4 series in preview on April 24, 2026, following DeepSeek-V3 in December 2024 and DeepSeek-R1 in January 2025, the two releases that made the lab internationally known. The series has two members: V4-Pro at 1.6 trillion total parameters with 49 billion activated, and V4-Flash at 284 billion total with 13 billion activated. Both support a one-million-token context and both were published as open weights under the MIT License.^[19][20]

At the same launch, DeepSeek pointed the two legacy API names, deepseek-chat and deepseek-reasoner, at V4-Flash in non-thinking and thinking mode respectively, and gave them a three-month retirement window ending July 24, 2026.^[1] That retirement took effect one week before the 0731 build shipped, so by the end of July the DeepSeek API exposed only two model ids: deepseek-v4-flash and deepseek-v4-pro.^[3]

Through the second quarter of 2026 the models on the API remained the April preview builds. DeepSeek's own July comparison chart labels them DeepSeek-V4-Flash-Preview and DeepSeek-V4-Pro-Preview.^[2]

What changed in the 0731 build

DeepSeek was unusually explicit about the scope of the update. According to the API change log, "DeepSeek-V4-Flash-0731 keeps the same model architecture and size as DeepSeek-V4-Flash-Preview, and was only re-post-trained."^[1] Nothing about the pre-trained backbone, the parameter count, the context length, or the price changed. The API calling method is unchanged as well: requests that already set model to deepseek-v4-flash picked up the new build automatically.^[1][9]

DeepSeek also scoped the update tightly in the other direction. The change log states that only the V4-Flash API was upgraded, and that the V4-Pro API and the models behind the DeepSeek app and website were left as they were, with the official V4-Pro release to "follow soon."^[1] Chinese-language coverage from IT之家 and TechNode reported the same split on the day.^[14][17]

Because the release is a post-training change rather than a new model, the practical effect for callers was a silent swap: the same model id returned different behavior from July 31 onward. That is worth knowing for anyone comparing evaluation runs across the boundary.

Benchmark results reported by DeepSeek

DeepSeek published nine agent-oriented scores in the change log and in a chart attached to its announcement post. The company's framing is that the new Flash build now beats the larger V4-Pro-Preview across the board.^[1][2]

Scroll sideways for more →

All of these are vendor-reported figures. Two footnotes on DeepSeek's own chart matter for reading them. First, the public code-agent tasks were run inside a scaffold DeepSeek calls the DeepSeek Harness in minimal mode, which had not been released at the time of the announcement, at max effort with top_p 0.95 and temperature 1.0.^[1][2] A model's score on agent benchmarks depends heavily on the harness driving it, so results measured with an unreleased first-party scaffold are not directly comparable to leaderboard entries produced with a public one. 36Kr noted that the announcement was the first time DeepSeek's in-house harness had appeared under an official name.^[16] Second, DSBench-FullStack and DSBench-Hard are DeepSeek's own internal test sets, described as a full-stack development set and a hard coding-agent set, so no outside party can reproduce those two rows.^[1]

What the benchmarks measure

Four of the public benchmarks have published documentation. Terminal-Bench 2.1 is an 89-task suite for coding agents working in a command-line environment, released on May 6, 2026 as a revision of Terminal-Bench 2.0 that repaired 28 tasks and added continuous validation of the task set.^[21] Toolathlon (the Tool Decathlon) covers 32 applications and 604 tools across 108 verified tasks that typically take around 20 turns, and its authors reported that the best model they tested reached 38.6 percent.^[22] CyberGym has 1,507 real vulnerabilities from 188 software projects and asks an agent to write a proof-of-concept test that reproduces each one from a text description plus source code.^[23] NL2Repo-Bench hands an agent a requirements document and an empty workspace and grades the Python library it produces against the original project's test suite, a setting in which the paper's authors found that even the strongest agents stayed below a 40 percent average pass rate.^[24] DeepSeek's rows are not directly comparable with the figures in those papers: its Toolathlon row is labelled "Toolathlon-Verified" rather than the full set, and every public code-agent run went through its own harness.

Independent evaluation

Artificial Analysis ran the model on the same day and published a breakdown that broadly supported DeepSeek's claim while trimming some of the numbers.^[10][11]

On the Artificial Analysis Intelligence Index v4.1, which aggregates nine evaluations (GDPval-AA v2, tau-cubed Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience and AA-LCR), DeepSeek V4 Flash 0731 in max reasoning effort scored 50, up from 40 for the April build and six points ahead of V4-Pro at 44.^[10][12] That put it level with Google's Gemini 3.6 Flash at 50 and one point behind GPT-5.6 Luna, Z.ai's GLM-5.2 and Meta's Muse Spark 1.1, all at 51. It remained seven points behind Kimi K3 from Moonshot AI at 57, which Artificial Analysis described as the open-weights frontier.^[10]

The agentic gains were the largest. Artificial Analysis measured GDPval-AA v2 at an Elo rating of 1559, up from 1189 for the preview, and its own Terminal-Bench 2.1 run at 79 percent, a 17-point rise but roughly four points below DeepSeek's reported 82.7. Its tau-cubed Banking score rose eight points to 31 percent. Every evaluation in the index improved: CritPt up nine points to 17 percent, SciCode up five to 50 percent, Humanity's Last Exam up five to 37 percent, AA-LCR up three to 66 percent and GPQA Diamond up one to 91 percent.^[10]

Two findings cut against a simple "it got smarter" reading. The AA-Omniscience result, an index score of -16 that Artificial Analysis called a 7-point improvement on the preview, came entirely from a lower hallucination rate rather than higher accuracy: the share of answers that were correct did not move, while the hallucination rate fell 12 points to 84 percent, which Artificial Analysis put in the same range as GPT-5.6 Terra at 85 percent.^[10] Separately, the model is verbose. It burned roughly 206 million output tokens as first reported on July 31 (Artificial Analysis's model page later showed 210 million) to complete the index, which is 12 percent fewer than the 234 million the April build needed but still far above the 62 million median across models.^[10][11]

The cost picture was the headline for Artificial Analysis. It placed V4 Flash 0731 on its Pareto frontier for intelligence against cost per task and calculated that, even after OpenAI cut GPT-5.6 Luna prices by 80 percent the day before, running the index on DeepSeek's first-party API cost about 60 percent less than on the comparably scoring OpenAI model. The firm attributed much of that to DeepSeek's roughly 98 percent cache-hit discount, against the 90 percent that most of the industry offers.^[10]

Responses API support and Codex compatibility

The 0731 release added a third wire protocol to the DeepSeek API, and for developers this may matter more than any of the benchmark rows. The Responses API is the request and response format OpenAI introduced for stateful, tool-using agentic work, and it is the protocol OpenAI's Codex clients speak. A provider that implements the same shapes can be dropped in behind a Codex install without patching the client. DeepSeek did exactly that at https://api.deepseek.com, and documents a ~/.codex/config.toml that sets wire_api = "responses" under a [model_providers.deepseek] block, plus a models.json catalog declaring the model's context window and reasoning levels to Codex.^[5][6] A one-line setup script is offered for macOS, Linux and Windows; DeepSeek says it backs up the existing configuration first and validates the files before writing.^[5] The same configuration serves the Codex CLI, the ChatGPT desktop app and the Codex extension for Visual Studio Code.^[5]

Support is currently limited to Flash. DeepSeek's documentation states in three places that the Responses API and the Codex integration cover deepseek-v4-flash only, with deepseek-v4-pro support expected in early August 2026.^[3][5][6]

Compatibility limits

DeepSeek's implementation is a stateless subset. previous_response_id and conversation are not supported, store always returns false, and background, metadata, include, prompt and service_tier are not implemented. Requests that exceed the context window return HTTP 400 rather than truncating, because truncation is unsupported. prompt_cache_key and prompt_cache_retention are not accepted either, since DeepSeek manages context caching automatically. Unsupported parameters are silently ignored so that existing clients connect without modification. Among tools, function and web_search work, custom is accepted only for the apply_patch name that Codex needs, and file_search, code_interpreter, computer_use and MCP tools are ignored. Image and file inputs are not supported: an input_image part does not raise an error but is replaced with placeholder text.^[6]

DeepSeek documents a separate integration for Anthropic's Claude Code through https://api.deepseek.com/anthropic, in which V4-Pro fills the Opus and Sonnet roles and V4-Flash is mapped to the Haiku and subagent roles.^[8]

Architecture and specifications

Because the 0731 build reuses the preview's network unchanged, its architecture is the V4 architecture described in DeepSeek's technical report, "DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence," submitted to arXiv on April 26, 2026.^[20] The three changes the paper foregrounds are a hybrid attention design that pairs Compressed Sparse Attention with Heavily Compressed Attention, Manifold-Constrained Hyper-Connections (mHC) in place of ordinary residual connections, and the Muon optimizer. The paper reports that at a one-million-token context, V4-Pro needs 27 percent of the per-token inference FLOPs and 10 percent of the KV cache of DeepSeek V3.2.^[19][20] Both models were pre-trained on more than 32 trillion tokens. The DeepSeek V4 article covers the architecture in detail.

Codex-facing metadata in DeepSeek's models.json fills in a few operational numbers that the model card does not: the declared context window is 1,048,576 tokens, Codex is told to treat 95 percent of that as effectively usable, parallel tool calls are always on, and the three reasoning levels are low ("Fast responses with lighter reasoning"), high ("Extra high reasoning depth for complex problems") and max ("Maximum reasoning depth for the hardest problems"), with high as the default.^[5]

Availability, pricing, and rate limits

V4-Flash is served from DeepSeek's first-party API at https://api.deepseek.com for OpenAI-format and Responses-format calls, and at https://api.deepseek.com/anthropic for Anthropic-format calls.^[3] Prices did not change with the 0731 build.^[3][10]

Scroll sideways for more →

DeepSeek's Chinese-language pricing page lists the same rates in yuan: 0.02, 1 and 2 yuan per million tokens for Flash, and 0.025, 3 and 6 yuan for Pro.^[4] The Pro figures reflect a price cut that BigGo Finance describes as a permanent reduction made in May 2026, from 24 yuan to 6 yuan per million output tokens.^[28] Both models are capped at 384K output tokens.^[3] Concurrency is counted per account across all API keys, and exceeding it returns HTTP 429; DeepSeek offers free capacity expansion on request and supports a user_id parameter for per-tenant KV-cache isolation, scheduling isolation and content-safety separation.^[7] For a broader view of how these rates sit against other vendors, see LLM API pricing comparison.

Peak and off-peak pricing

A footnote on the pricing page states that DeepSeek "will soon adopt a peak/off-peak pricing policy," under which prices during peak hours will be twice the regular prices and the change will apply to all billing items. Peak hours are defined as 9:00 to 12:00 and 14:00 to 18:00 Beijing time (UTC+8) daily, and the effective date is left to a future official announcement.^[3] The identical note appears on the Chinese version of the page.^[4] As of July 31, 2026 the policy had not taken effect and no start date had been published, so the rates in the table above are the ones actually charged.

Doubling would put V4-Flash output at 4 yuan per million tokens and V4-Pro output at 12 yuan during working hours.^[28] The scheme reverses the direction of DeepSeek's earlier experiment with time-based pricing: from February 26, 2025 the company ran off-peak discounts instead, cutting DeepSeek-V3 by 50 percent and DeepSeek-R1 by 75 percent between 16:30 and 00:30 UTC, a program it wound down the following September.^[27] BigGo Finance reported that Chinese developers objected that the proposed peak window covers almost the whole working day in their time zone, effectively favoring users in other regions, while others argued the doubled rate would still undercut competing APIs.^[28]

Weights and licensing

The 0731 build is API-only at launch. As of July 31, 2026 the deepseek-ai/DeepSeek-V4-Flash repository on Hugging Face still carried the April checkpoint and opened with the sentence "We present a preview version of DeepSeek-V4 series," with no 0731 revision published.^[19] Artificial Analysis wrote that DeepSeek "is expected to release the model's full weights in the coming weeks," and noted that on its GDPval-AA v2 result the model would then rank second among open-weights models behind Kimi K3.^[10] That expectation is a forecast, not a DeepSeek commitment; the company has not published a date.

The preview weights remain available under the MIT License in FP4-plus-FP8 mixed precision, along with a base checkpoint, and are mirrored on ModelScope.^[19] Third-party hosts carry the earlier build: OpenRouter, for example, lists DeepSeek V4 Flash with a one-million-token context at $0.09 per million input tokens and $0.18 per million output tokens through routed providers.^[25] For readers comparing terms across labs, see open-weight license comparison and open source AI.

Reception and adoption

Coverage on launch day concentrated on the inversion of DeepSeek's own product ladder: a 13-billion-active-parameter model outscoring the lab's 49-billion-active flagship preview on agent work. The Chinese outlet 智东西 ran the story under a headline calling it "倒反天罡" (roughly, the natural order turned upside down) and argued that the release shows low price does not imply low capability.^[18] 36Kr reported DeepSeek's framing that for specific tasks, model scale is not the decisive factor once post-training optimization is applied properly.^[16] BigGo Finance reported that analysts read the Flash-first sequencing as aimed at enterprise demand for low-cost, high-concurrency automation.^[15] OfficeChai's write-up compared the reported scores against Opus 4.8 and against GPT-5.6 Sol, Claude Fable 5 and Claude Sonnet 5 list prices, all of which V4-Flash undercuts by a wide margin.^[13]

Third-party coding agent products moved quickly. The agent tool Command Code announced support the same morning, saying the model was live on all of its plans with tool calls and cache repairs.^[26]

Limitations

The model is text-only in and text-only out; DeepSeek's Responses API documentation confirms that image and file inputs are dropped.^[6][10] The Responses implementation is stateless, so applications that rely on server-side conversation threading through previous_response_id need to manage history themselves.^[6] Verbosity is a real cost driver despite the low per-token price: Artificial Analysis measured roughly 206 million output tokens to complete its index, more than three times the median.^[10] The AA-Omniscience result shows an 84 percent hallucination rate, improved but still high, and factual accuracy did not move between builds.^[10] Vendor benchmark numbers were produced with an unreleased in-house harness, two of the nine reported benchmarks are internal and unreproducible, and the one directly comparable public figure, Terminal-Bench 2.1, came in about four points lower when measured independently.^[1][10] Finally, the API is in public beta, the matching V4-Pro build and Codex support were still pending as of July 31, and no weights had been published for the 0731 model.^[1][5][19]

Significance

The 0731 release is a data point about where capability gains were coming from in mid-2026. DeepSeek changed no architecture, added no parameters and raised no prices, yet the model moved from 40 to 50 on an independent aggregate index and from 7.3 to 54.4 on one vendor-reported coding benchmark. DeepSeek attributes the change entirely to redone post-training, and reported its public code-agent scores using a first-party harness, so part of the measured gain belongs to the scaffold rather than the weights. The lab has not published a technical description of the 0731 post-training run.^[1]

The protocol work carries its own weight. By implementing OpenAI's Responses API and publishing a Codex configuration, DeepSeek made its model a drop-in backend for tooling that developers outside China already run, which lowers switching costs in a way a benchmark table does not. Paired with a 98 percent cache-hit discount and per-token prices roughly an order of magnitude below comparably scoring closed models, that combination, rather than a frontier score, is the competitive position the release stakes out.^[5][10]

See also

• DeepSeek V4
• DeepSeek
• DeepSeek V3.2
• OpenAI Codex
• OpenAI Responses API
• Kimi K3
• Context caching
• LLM API pricing comparison

References

1. ^Change Log (entry dated 2026-07-31), DeepSeek API Docs
2. ^DeepSeek-V4-Flash Official API is now LIVE in public beta, @deepseek_ai on X, July 31, 2026
3. ^Models & Pricing, DeepSeek API Docs (accessed July 31, 2026)
4. ^模型 & 价格, DeepSeek API 文档 (accessed July 31, 2026)
5. ^Integrate with Codex, DeepSeek API Docs
6. ^Using the Responses API, DeepSeek API Docs
7. ^Rate Limit & Isolation, DeepSeek API Docs
8. ^Integrate with Claude Code, DeepSeek API Docs
9. ^Your First API Call, DeepSeek API Docs
10. ^DeepSeek V4 Flash 0731 scores 50 on the Artificial Analysis Intelligence Index, @ArtificialAnlys on X, July 31, 2026
11. ^DeepSeek V4 Flash: Intelligence, Performance & Price Analysis, Artificial Analysis
12. ^DeepSeek-v4-Flash-0731 Scores 50 On Artificial Analysis Intelligence Index, Creates Big Spike On Pareto Frontier, OfficeChai, July 31, 2026
13. ^DeepSeek Releases DeepSeek-V4-Flash-0731, Gives Opus 4.8-Level Performance At A Fraction Of The Price, OfficeChai, July 31, 2026
14. ^DeepSeek puts V4-Flash API into public beta, TechNode, July 31, 2026
15. ^DeepSeek Launches V4-Flash API Public Beta with Major Agent Capability Leap, BigGo Finance, July 31, 2026
16. ^DeepSeek V4 Official Version Released, 36Kr, July 31, 2026
17. ^DeepSeek-V4-Flash 正式版 API 上线公测，V4-Pro 正式版将"尽快"发布, IT之家, July 31, 2026
18. ^倒反天罡？DeepSeek-V4-Flash 正式版上线，干翻自家先发的 Pro 版, 智东西, July 31, 2026
19. ^deepseek-ai/DeepSeek-V4-Flash model card, Hugging Face
20. ^DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence, arXiv:2606.19348 (April 26, 2026)
21. ^Terminal-Bench 2.1, tbench.ai (May 6, 2026)
22. ^The Tool Decathlon: Benchmarking Language Agents for Diverse, Realistic, and Long-Horizon Task Execution, arXiv:2510.25726
23. ^CyberGym: Evaluating AI Agents' Real-World Cybersecurity Capabilities at Scale, arXiv:2506.02548
24. ^NL2Repo-Bench: Towards Long-Horizon Repository Generation Evaluation of Coding Agents, arXiv:2512.12730
25. ^DeepSeek: DeepSeek V4 Flash, OpenRouter
26. ^DeepSeek V4 Flash 0731 now live in Command Code, @CommandCodeAI on X, July 31, 2026
27. ^Off-Peak Discounts Alert, @deepseek_ai on X, February 26, 2025
28. ^DeepSeek to Introduce Peak/Off-Peak Pricing for V4 Official Release, Doubling API Costs During Peak Hours and Sparking Developer Debate, BigGo Finance