TA/deepseek-ai/DeepSeek-V3

Common Name: DeepSeek V3

TogetherAI
Released on Feb 17, 2025 12:00 AM
CompareTry in Chat

DeepSeek V3 MoE model with 671B total parameters and 37B active, hosted on TogetherAI.

Specifications

Context
64K
Maximum Output
8K
Inputtext
Outputtext

Performance (7-day Average)

Collecting…
Collecting…
Collecting…

Pricing

Input$1.38/MTokens
Output$1.38/MTokens

Availability Trend (24h)

Performance Metrics (24h)

Similar Models

$0.97/$0.97/M
ctx128Kmaxavailtps
InOut

Meta's Llama 3.1 70B optimized for fast inference on TogetherAI.

$1.32/$1.32/M
ctx128Kmaxavailtps
InOut

Alibaba's Qwen2.5 7B model optimized for fast inference on TogetherAI.

$0.17/$0.66/M
ctx128Kmaxavailtps
InOutCap

OpenAI's open-weight 120B model for production and high reasoning use cases, hosted on TogetherAI.

$0.06/$0.22/M
ctx128Kmaxavailtps
InOutCap

OpenAI's open-weight 20B model for lower latency and local use cases, hosted on TogetherAI.