Command Palette
Search for a command to run...
tokenspeed
TokenSpeed is a lightweight Python-based inference engine designed to maximize throughput and minimize latency for large language models
LivenessActiveRankScore 27.7Last commit—SO tag heat—Press sentiment 7d—Hiring—Cohort 2026+36.1 WoW
What it is
TokenSpeed is a lightweight Python-based inference engine designed to maximize throughput and minimize latency for large language models. It optimizes token generation speed across popular LLM architectures including Deepseek, Qwen, and Kimi, enabling faster real-time inference at scale.
Signal score
vs category avg: +-37.327.7+36.1 WoW
Funding history
1 roundTotal raised
$2.0B
Latest valuation
$20.0B
Stage
unknown
Lead investors: Meituan
Data from SEC EDGAR Form D filings and press coverage. Round labels for SEC filings are inferred from filing sequence and amount.
Moat
Defensibility classification confidence below display threshold (0.7). Re-evaluating.
Press mentions
0 lifetime · 0 shownNo press mentions found.
Discussion 0 comments
No comments yet. Be the first!
Vitalslive · 05:55 UTC
Signal
27.7+36.1
Rank in cat
—
Upvotes
—
Press 30d
—
Founders
Not yet researched. Weekly founder enrichment runs Sundays.
Liveness probedaily 04:30 UTC
Active · 200 OK
lightseek.org · last 13:42 UTC
30-day probe history
30d agotoday
GitHub activitydaily 04:15 UTC
Last commit
—
Last release
—
Open PRs
—
Open issues
32
Contributors 30d
—
Closed 30d
—
Stack Overflow tag heatweekly Sun
Category tags · 30d / lifetime ratio
0.59%
0.035%
0.14%
0.057%
Category heat (avg)—
Press sentiment · 7d rollinghourly
+0.72
Strongly positive
7 mentions in window
growthagenticproductvs competitors
Cohort survival · Class of 2026weekly
Products in the 2026 cohort. Cohort survival data is not yet computed for individual products.
Median Lifespan
4.2y
Cohort Rank
—