Skip to content

Meta vs DeepSeek

as of

Meta and DeepSeek side by side on everything we track. Each lab’s own page has the full detail.

Readout

Both labs at a glance. Every figure links to the lab’s page, which cites its source.

Each lab’s flagship price, disclosed revenue, demand share, open roles, incidents, and newest alert
Lab Flagship $/Mtok SEC revenue Demand share Open roles Incidents 30d Newest alert
Meta Tear sheet $4.00 / $20.00in / out · muse-wonder-1.0None filed1.3%−0.1 ppof all OpenRouter tokens, 7 daysNo public feedNo public feednone
DeepSeek Tear sheet Not mappedNone filed39.2%+3.5 ppof all OpenRouter tokens, 7 daysNo public feed7vs 8none opennone

API prices

$ per million tokens, like for like on each vendor’s own recommendation.

Like-for-like API pricing in dollars per million tokens, by lab and class
Lab Flagship the vendor’s own “start here” model for hard workBalanced the vendor’s stated intelligence-for-cost tradeoffEconomy the vendor’s model for cost-sensitive, high-volume work
Meta
muse-wonder-1.0

$4.00 in $20.00 out

dev.meta.ai
Not mapped
muse-spark-1.3

$1.25 in $4.25 out

dev.meta.ai
DeepSeek
Not mapped
Not mapped
Not mapped

Each lab in full: Meta, DeepSeek.

Data from the last 90 days is CC BY 4.0. Older data is CC BY-NC 4.0, except what we published before 2026-11-04, which stays CC BY 4.0. How you can reuse it.