DeepSeek V3.1

DeepSeektext → textHugging Face OpenRouter

The cheapest provider for DeepSeek V3.1 is DeepInfra at $0.425 per 1M tokens (blended), charging $0.25 per 1M input and $0.95 per 1M output tokens. 7 providers serve it; the most expensive charges 2.1× as much. Context window: 160K tokens. Prices tracked since Oct 3, 2026.

Best price

$0.425

at DeepInfra

Input / output

$0.25 / $0.95

per 1M tokens

Providers

2.1× spread

7 days

—

best price

30 days

—

best price

Context

160K

tokens

Providers

Active offers ranked by blended price. Bars are relative to the most expensive.

Provider prices for DeepSeek: DeepSeek V3.1, USD per 1M tokens
Provider	Input	Output	Blended	vs cheapest	Context	Uptime 1d	Price since
DeepInfraCheapest fp4	$0.25	$0.95	$0.425	best	160K	99.9%	Oct 3
SiliconFlow fp8	$0.27	$1.00	$0.453	+6%	160K	98.1%	Oct 3
AtlasCloud fp8	$0.30	$1.00	$0.475	+12%	128K	99.5%	Oct 3
CoreWeave fp8	$0.55	$1.65	$0.825	+94%	161K	100%	Oct 3
SambaNova fp8	$0.65	$1.50	$0.863	+103%	128K	96.3%	Oct 3
Googleus-west2	$0.60	$1.70	$0.875	+106%	160K	0.0%	Oct 3
Mara	$0.60	$1.70	$0.875	+106%	128K	99.4%	Oct 3

DeepInfraCheapest
$0.425
$0.25 in · $0.95 out
best
160K ctx99.9% uptimefp4
SiliconFlow
$0.453
$0.27 in · $1.00 out
+6%
160K ctx98.1% uptimefp8
AtlasCloud
$0.475
$0.30 in · $1.00 out
+12%
128K ctx99.5% uptimefp8
CoreWeave
$0.825
$0.55 in · $1.65 out
+94%
161K ctx100% uptimefp8
SambaNova
$0.863
$0.65 in · $1.50 out
+103%
128K ctx96.3% uptimefp8
Google
us-west2
$0.875
$0.60 in · $1.70 out
+106%
160K ctx0.0% uptime
Mara
$0.875
$0.60 in · $1.70 out
+106%
128K ctx99.4% uptime

Blended = (3 × input + output) / 4.

Price history

Hover to compare providers at any point in time. Click a legend item to hide it.

USD per 1M tokens · each step is a price change

Not enough history yet

Tracking since Oct 3, 2026. The chart fills in as new crawls arrive every 6 hours.

Other pricing

Cache, reasoning and per-call fees. Token rates per 1M; others per unit.

Offered by	Cache read
DeepInfra	$0.13
AtlasCloud	$0.135
CoreWeave	$0.55

Change log

Every recorded change for this model since Oct 3, 2026.

No changes yet

Tracking since Oct 3, 2026. Price changes, new providers and delistings will show up here.

About this model

DeepSeek-V3.1 is a large hybrid reasoning model (671B parameters, 37B active) that supports both thinking and non-thinking modes via prompt templates. It extends the DeepSeek-V3 base with a two-phase long-context...

Released: Aug 21, 2025
First tracked: Oct 3, 2026
Canonical slug: deepseek/deepseek-chat-v3.1
Weights: deepseek-ai/DeepSeek-V3.1