Polymarket has opened sandbox testing for Nex, a trading system rewritten in Rust that its engineering team says can handle 100,000 commands per second, ahead of a conditional mid-December launch.
Summary
- Nex targets 100,000 commands per second at launch, with capacity for short bursts of 150,000.
- Internal p99 order placement fell from 1.55 seconds to 3.76 milliseconds in company testing.
- Four weeks of shadow testing will precede deployment, while WebSocket improvements could arrive sooner.
- Existing requests and responses will remain compatible, according to Polymarket’s engineering announcement.
Polymarket’s vice president of engineering for DeFi, Josh Stevens, announced the sandbox on X on Nov. 13, saying Nex had spent the previous week running alongside the existing exchange. It mirrors every production order so engineers can compare response times, stability under load, and matching accuracy against live trading activity.
https://x.com/devjoshstevens/status/2121241806469796103
The exchange has not switched over to Nex. Stevens said the team plans four weeks of shadow operation before deciding whether to deploy it in mid-December.
Polymarket Nex targets 10 times the current capacity
In the performance figures Stevens published, the current central limit order book, or CLOB, handles 10,000 commands per second under steady load and 15,000 during short bursts. Nex supports 100,000 and 150,000, respectively.
Commands cover actions such as placing and canceling orders; the figures do not describe completed trades per second.
With further optimization, Stevens said the system could reach 200,000 commands per second, with bursts of 250,000. More extensive changes to its underlying code could take capacity to 1 million commands per second over the next year, although he said that level was not yet needed.
Startup time has fallen from five minutes to 15 seconds, according to the announcement. Stevens described the old CLOB as having race conditions and inconsistent behavior, saying Nex was designed to produce consistent results and remove those problems.
The team began the rewrite in August, after concluding that the existing system needed to be rebuilt from scratch.
Order benchmarks measure internal exchange speed
At the 99th percentile, or p99, order placement took 3.76 milliseconds on Nex against 1.55 seconds on the existing system, a roughly 412-fold improvement in the published comparison. P99 measures the time within which 99% of measured requests finish.
Canceling orders fell from 627.92 milliseconds to 3.20 milliseconds, while canceling market orders dropped from 810.16 milliseconds to 5.14 milliseconds.
“The numbers below are internal exchange timings, not e2e,” Stevens wrote, referring to end-to-end performance. He said those measurements would follow.
The distinction means the published figures do not measure the full journey from a trader submitting an instruction to receiving the result.
At p95, order placement fell from 341.84 milliseconds to 1.59 milliseconds. The corresponding figures were 78.22 milliseconds to 1.58 milliseconds for canceling orders, and 84.89 milliseconds to 2.27 milliseconds for canceling market orders.
Cancellation and market feeds get separate changes
The announcement described a change in how the exchange handles requests to cancel every open order. The current CLOB acknowledges the request before finishing the cancellations in the background, usually taking another 600 milliseconds to two seconds.
Nex responds after every order has actually been canceled, Stevens said. Its p99 completion time was 3.14 milliseconds, compared with approximately 0.64 to 2.04 seconds for the existing system’s full cancellation process.
The rewritten Rust WebSocket service, which delivers order-book changes, prices and fills, recorded p99 write latency of 0.099 milliseconds against 44.5 to 88.1 milliseconds previously. Stevens said delayed feeds can leave traders quoting against old prices.
WebSocket improvements can be deployed separately from the exchange, so the team may release them before Nex goes live.
Developers are promised compatible requests and responses
Stevens said Nex introduces no breaking changes to requests or responses for market makers and builders. The sandbox lets participants compare it with the current CLOB and submit feedback before deployment.
The work follows Polymarket’s in-house indexing overhaul, which crypto.news covered in September. Stevens said that the Rust-based system surfaced some blockchain events up to 14 blocks, or roughly 28 seconds, faster, with Goldsky retained as backup.
In October, Polymarket outlined separate protocol changes involving position tokens, collateral, and market resolution, with Nov. 2 then set as a tentative transition date for newly created markets.
The company’s international trading operation remains separate from its regulated U.S. venue. Stevens’ Nex announcement specifically concerns Polymarket DeFi.
The team left some optimizations out of the initial release to get Nex ready sooner. Stevens said work on those changes would begin in the first quarter of next year, with a target of making p95 and p99 timings another five times faster.
