DeepSeek-V4-Flash-0731

Released

DeepSeek-V4-Flash-0731 is an AI model released by DeepSeek on Friday, Jul 31 2026, 98 days after DeepSeek-V4-Flash. Benchmark results (shown below) cover BullshitBench v2, SWE-Bench Verified, Terminal-Bench 2.1, Toolathlon-Verified, BrowseComp, CyberGym, and 2 more.

API pricing

Off-peak (all hours outside the published peak windows)
Input
$0.22
Cached input
$0.007
Output
$0.66
Peak (01:00–04:00 and 06:00–10:00 UTC)
Input
$0.44
Cached input
$0.014
Output
$1.32
  • Thinking and non-thinking modes use the same published price.
USD per 1M tokens

Available from

OpenInference$0.04$0.101Mfp8
Baidu$0.0352$0.10561Mfp8
Relace$0.06$0.121Mfp4
Inceptron$0.0619$0.15211Mfp4
StreamLake$0.0572$0.17161Mfp8
DeepInfra$0.06$0.181Mfp8
Makora$0.09$0.1951M
Waferfast$0.10$0.251M
DigitalOcean$0.08$0.2521M
BaseTen$0.13$0.261Mfp8
CoreWeave$0.13$0.28256Kfp8
Parasail$0.14$0.281Mfp8
Together$0.14$0.281M
Morph$0.1234$0.34751Mbf16
Venice$0.175$0.351M
Sail Research$0.078$0.361Mfp4
Mancer 2$0.20$0.601Mfp8
Reka$0.11$0.66256Kfp4
Fireworks$0.22$0.661M
GMICloud$0.286$0.8581Mfp8
Alibaba$0.352$1.0561M
NextBit$0.352$1.0561Mfp8
Novita$0.4092$1.22761Mfp8
AtlasCloud$0.44$1.321Mfp4
Cloudflare$0.44$1.321.3M
Phala$0.44$1.321M
USD per 1M tokens

Benchmarks

Coding

SWE-Bench Verified
79%
#16 of 57

Terminal & CLI

Terminal-Bench 2.1
82.7%
#15 of 30

Agentic & tool use

Toolathlon-Verified
70.3%
#8 of 9
BrowseComp
73.2%
#22 of 28
CyberGym
76.7%
#9 of 10
AutomationBench
25.1%
#11 of 11

Reasoning & science

Humanity's Last Exam
34.8%
with tools
#32 of 42

Robustness

BullshitBench v2
39%
#37 of 77

Source: BenchLM, retrieved 31 August 2026. Every other score here is the figure the lab published at launch.

About

DeepSeek-V4-Flash-0731, announced July 31, 2026 as the "official" V4-Flash API in public beta, was a post-training refresh of the April V4-Flash build — same architecture, retrained for agentic work. The jump was unusually large for a checkpoint update: on DeepSeek's own evaluation harness it scored 82.7 on Terminal-Bench 2.1 against 61.8 for the April build, and 76.7 on CyberGym, surpassing the larger V4-Pro-Preview on several coding benchmarks despite being the cheaper tier. DeepSeek's harness numbers for competitor models differed from those labs' own published figures, so its launch chart is best read as within-family comparison.

The release doubled as a renaming: DeepSeek retroactively designated the April 24 builds V4-Flash-Preview and V4-Pro-Preview, and served the new checkpoint under the unchanged deepseek-v4-flash API name — existing callers were switched to it without a code change. It added native support for the OpenAI Responses API format and Codex compatibility. Unlike every prior DeepSeek release, it shipped API-only at launch: no 0731 weights were published, and Hugging Face still carried the April build at release. V4-Pro, the app, and the web product stayed on their April checkpoints until the 0813 Pro refresh two weeks later. The Flash tier moved off the V4 architecture entirely with DeepSeek-V4.1-Flash in September 2026.

Compare DeepSeek-V4-Flash-0731 with

DeepSeek-V4-Flash-0731

Suggested comparisons

Frequently asked questions

DeepSeek-V4-Flash-0731 was released by DeepSeek on Friday, Jul 31 2026.

All DeepSeek releases

23 tracked