DeepSeek: DeepSeek V4 Pro 0423

deepseek/deepseek-v4-pro

DeepSeek V4 Pro 0423 is an open-weight model from DeepSeek, ranking #68 of 235 on the TryAii Score. It's strongest on knowledge (#14 on GPQA). It's served by 21 independent hosts from $0.992/$1.98 per million input/output tokens — a little under the going rate for its score class.

Prompt $/1M

$0.992

Completion $/1M

$1.984

Context

1,048,576 tok

Best provider

80 tok/s · Cloudflare

Specs

Model id: deepseek/deepseek-v4-pro
Display name: DeepSeek: DeepSeek V4 Pro 0423
Modality: text->text
Tokenizer: DeepSeek
Instruct type: —
Context length: 1,048,576
Max completion: 384,000
Prompt $/1M: $0.992
Completion $/1M: $1.984
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
Cloudflare	80	2,223 ms	4,774 ms
BaseTen	69	543 ms	660 ms
Artificial Analysis	67	1,080 ms	—
Baidu	61	808 ms	923 ms
CoreWeave	61	426 ms	519 ms
Parasail	60	734 ms	1,028 ms
Fireworks	57	742 ms	964 ms
Novita	53	1,595 ms	1,813 ms
Alibaba	50	1,249 ms	1,352 ms
NextBit	50	2,414 ms	2,722 ms
AtlasCloud	45	1,521 ms	2,447 ms
Azure	45	1,480 ms	2,184 ms
Together	43	939 ms	2,071 ms
SiliconFlow	36	2,413 ms	2,917 ms
DeepSeek	35	1,135 ms	1,296 ms
StreamLake	34	2,486 ms	3,006 ms
Wafer	33	6,169 ms	7,030 ms
Ionstream	28	805 ms	1,148 ms
DigitalOcean	27	930 ms	1,204 ms
Venice	26	3,107 ms	4,625 ms
GMICloud	25	3,315 ms	4,575 ms
DeepInfra	22	1,254 ms	1,988 ms
Io Net	13	968 ms	968 ms
WandB	7	1,407 ms	1,695 ms
deepseekfastest	—	—	—