DeepSeek: DeepSeek V4.1 Flash

deepseek/deepseek-v4.1-flash

DeepSeek V4.1 Flash is an open-weight model from DeepSeek, ranking #61 of 266 on the TryAii Score. It's strongest on head-to-head chat (#18 on Chatbot Arena Elo (Code)). It's served by 29 independent hosts from $0.02/$1 per million input/output tokens — about a third of what models at its level usually charge.

Prompt $/1M

$0.020

Completion $/1M

$1.000

Context

1,048,576 tok

Best provider

280 tok/s · Together

Specs

Model id: deepseek/deepseek-v4.1-flash
Display name: DeepSeek: DeepSeek V4.1 Flash
Modality: text+image->text
Tokenizer: DeepSeek
Instruct type: —
Context length: 1,048,576
Max completion: 943,718
Prompt $/1M: $0.020
Completion $/1M: $1.000
Free tier: no

Capabilities

Function callingno
JSON modeno
System promptno
Visionno
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
Together	280	186 ms	244 ms
Venice	234	317 ms	406 ms
Wafer	220	308 ms	348 ms
BaseTen	217	286 ms	297 ms
Fireworks	204	240 ms	284 ms
DigitalOcean	195	585 ms	718 ms
Cohere	189	283 ms	325 ms
PrimeIntellect	168	324 ms	357 ms
Relace	161	463 ms	711 ms
Parasail	154	338 ms	455 ms
Makora	142	237 ms	291 ms
InferenceNet	138	283 ms	369 ms
Novita	133	922 ms	986 ms
Crusoe	125	282 ms	386 ms
CoreWeave	122	1,012 ms	1,444 ms
Modal	118	390 ms	463 ms
ModelRun	117	216 ms	261 ms
Baidu	116	774 ms	919 ms
AtlasCloud	105	809 ms	867 ms
Reka	96	301 ms	366 ms
Morph	95	615 ms	810 ms
OpenInference	93	952 ms	1,228 ms
Decart	93	649 ms	709 ms
GMICloud	88	2,058 ms	2,256 ms
Nebius	84	1,458 ms	1,597 ms
Alibaba	69	1,135 ms	1,359 ms
SiliconFlow	69	1,810 ms	2,107 ms
DeepInfra	64	906 ms	1,349 ms
Phala	63	1,080 ms	1,157 ms
StreamLake	60	2,597 ms	3,201 ms
Sail Research	59	515 ms	738 ms
Google	43	790 ms	1,140 ms
Mancer 2	34	807 ms	899 ms