DeepSeek: DeepSeek V4 Flash 0731

deepseek/deepseek-v4-flash-0731

DeepSeek V4 Flash 0731 (DeepSeek) is a low-cost open-weight model priced near the bottom of the board, at $0.119/$0.238 per million input/output tokens, served by 21 independent hosts. For the money it delivers #121 of 268 on the TryAii Score. Its best showing is #20 on LiveCodeBench. It's reasonably quick at roughly 38-110 tokens/sec.

Prompt $/1M

$0.119

Completion $/1M

$0.238

Context

1,048,576 tok

Best provider

181 tok/s · Cohere

Specs

Model id: deepseek/deepseek-v4-flash-0731
Display name: DeepSeek: DeepSeek V4 Flash 0731
Modality: text->text
Tokenizer: DeepSeek
Instruct type: —
Context length: 1,048,576
Max completion: 943,718
Prompt $/1M: $0.119
Completion $/1M: $0.238
Free tier: no

Capabilities

Function callingno
JSON modeno
System promptno
Visionyes
Moderatedno

Benchmark scores

Each axis is one benchmark, normalized to the best model in the field.

No benchmark data for this model.

Speed by provider

Provider	tok/s	TTFT	Total latency
Coherefastest	181	401 ms	481 ms
Wafer	142	327 ms	356 ms
Venice	118	536 ms	663 ms
BaseTen	118	295 ms	384 ms
CoreWeave	115	418 ms	495 ms
Alibaba	111	1,678 ms	1,815 ms
DigitalOcean	107	827 ms	989 ms
Relace	69	545 ms	738 ms
Novita	69	1,167 ms	1,454 ms
Baidu	65	1,375 ms	1,502 ms
Reka	64	808 ms	1,203 ms
Cloudflare	64	1,136 ms	1,531 ms
Phala	60	2,109 ms	2,403 ms
AtlasCloud	57	1,517 ms	1,951 ms
GMICloud	52	2,344 ms	2,804 ms
Parasail	40	774 ms	942 ms
Inceptron	38	1,021 ms	1,458 ms
Mancer 2	35	3,181 ms	4,500 ms
StreamLake	34	3,366 ms	4,009 ms
DeepInfra	27	2,262 ms	3,171 ms
SiliconFlow	20	2,740 ms	3,407 ms
Sail Research	16	7,029 ms	7,029 ms