DeepSeek V3 0324 deepseek-chat-v3-0324

deepseek-ai · deepseek

unranked — no quality signals yet

params
684.5B
arch
moe
context
160k
license
mit open weights
released
Mar 2025
downloads/30d
631.7k

# architecture

attn shared MHA ×128 router top-8 of 256 +1 shared expert ×256 expert out ×61 layers ctx 163,840 684.5B pool
layers
61
d_model
7168
heads
128q / 128kv
head dim
56
experts
top-8 of 256 + 1 shared
vocab
129280
family
deepseek_v3

signal path of one layer, generated from the registry's structured fields — dimension lines quote the model's real numbers; an MoE trace forks at the router, a dense trace runs straight through. Geometry fields come from the repo's config.json.

# trend

collecting history — trends appear after a few daily runs.

# where to run

providerquantctx$/M in$/M out$/M cacheprice srctpsuptime✓
Featherless — ——— via hfrouter — —
DeepInfra 160k $0.24$0.90— via hfrouter 69 —
Gmi 160k $0.28$0.88— via litellm — —
SiliconFlowfp8 160k $0.25$1.00— via openrouter — 100.0%
GMICloudfp8 125k $0.29$1.14$0.11 via openrouter — 98.3%
Nebius 125k $0.50$1.50— via litellm — —
BaseTen — $0.77$0.77— via litellm — —
Hyperbolic 160k $1.25$1.25— hyperbolic — — ✓
Crusoe 160k $1.50$1.50— via litellm — —

sorted by blended price ((3·input + output) / 4 per 1M) · ✓ = the provider's own catalog confirms the offer · "via …" prices are what the aggregator routing the offer charges, not the provider's own list price

source aliases
openrouter
deepseek/deepseek-chat-v3-0324