
Tetris with real-time Jev decisions (X)
A video demo of Jev playing Tetris in real time, with the model's response speed fast enough to push blocks down during play.
Entries 49 to 79 of 79, in the hub's default order. The introduction is on page 1.
Showing 49 to 79 of 79

A video demo of Jev playing Tetris in real time, with the model's response speed fast enough to push blocks down during play.

Video Tetris benchmark: Laya-MLX runs fast on a MacBook at about 84ms per decision but loses to Jev in all three rounds.
73k viewsLaya latency ~84 msResult Jev won 3/3

A video demo benchmarking Jev running 500 real-time agents in parallel inside a 3D environment, reporting 500ms average latency and 35 API calls per second with an unoptimized, naive implementation.
61k viewsAgents 500Avg latency 500 ms

A video demo pits Jev against GLM 5.3 at chess: GLM won by checkmate in 29 moves, with Jev responding in about 0.3 seconds per move versus GLM's 5.8 seconds, for a total game cost of 24 cents.
39k viewsJev per move ~0.3sGame cost 24 cents

30-second Snake benchmark: a local 421M Laya checkpoint scores 46 at 86.5 decisions per second, while Jev 1.13.0 over the cloud API scores 1 at 3.2 decisions per second.
24k viewsLaya speed 86.5/sJev speed 3.2/s

Video benchmark where Jev wins outright as a cloud model, while a self-built JAX model ties with Laya-MLX among the local, on-device options.

A live first person shooter arena where Jev fights OpenJev, an open weights reimplementation running Qwen on SGLang. Matches are shareable by seed and run fresh each time rather than replaying a recording.

A browser tic-tac-toe game where Jev picks its move each turn with the choice primitive over the open squares. A move journal shows the probability Jev gave each square, a blitz mode gives the human one second per move, and every finished round joins a shared scoreboard.

Video demo from Inception's Mercury Decide launch post, showing it playing Jev at chess. Inception claims up to 14 decisions per second and the top JevBench v1.4 spot on OpenRouter; neither was independently run.

Open source snake game where two Jev agents compete against each other in real time to reach the trophy first, built to show off Jev's low latency in a game that needs real-time input.

Mario Never Dies, built with Jev and microsandbox: Jev picks every move, and each time Mario dies the VM forks into 4 timelines that keep playing; whichever survives becomes canon.

A video demo simulates a 120-person town where Jev decides each resident's actions every hour, plus town-level variables such as traffic, prices, and tax rate, producing over 20,000 decisions in under 5 minutes at a fraction of a cent.

Custom Snake benchmark comparing Jev over the API against locally run Laya-MLX and Kev-4B: Jev had zero deaths, Kev died twice at about 175ms latency, and Laya died early despite roughly 33ms latency.

A live Minesweeper broadcast where Jev picks every reveal and flag with the choice primitive and everyone watches the same game. A decision log shows each move's confidence and alternatives, a separate checker labels each loss as a forced guess or a missed deduction, and the game pauses when nobody is watching.

After close to 24 hours of reinforcement learning training on Doom, the author's Laya model still falls well short of Jev's zero-shot performance.

A video benchmark where local Laya beats API-hosted Jev 43-1 in Snake with wifi off: 86.4 decisions per second at 9.2ms P50 latency for Laya against 3.2 decisions per second at 308.7ms round trip for Jev, same schema and seed.

Community experiment driving the adventure game Kyrandia 2 with Jev choices; the model ends up looping. A useful failure case for real-time control.
A small village streamed live to the browser, where five families forage, farm, trade, gossip and raise children and dogs. Every judgment call a villager makes comes from Jev, and a second small model keeps a running newspaper on what happens.

Video of a Unity scene running Laya locally as a tic-tac-toe opponent, using receptron's ONNX export of the model through asus4's onnxruntime-unity plugin.

Head-to-head claim that Jev beats Claude Opus 5 at move selection in Pokemon Showdown.

A video of Jev and Laya fighting a 1v1 Doom deathmatch with monsters between them, first to five kills wins. Jev's calls were sharper and needed fewer corrections, but Laya decided twice as often and won 5 to 1.

Built an arena survival game where every enemy's move is decided live by Jev or Laya, turning each moment into one typed question at about 30 decisions a second. Laya ran roughly 21ms per decision on a laptop for no cost, while Jev was more accurate out of the box on the poster's own benchmark.

Video demo of Pac-Man controlled entirely by typed decisions from Laya-MLX running locally on Apple Silicon: 351 steps, a win, and 224 guardrail interventions logged.
An Overcooked-style kitchen simulation with four Jev-driven chefs handling expiring orders and competing priorities. In one early test Jev gave firefighting a 66% probability over finishing an order at 33%, then returned its move in about 200ms.

Video demo of a harness for playing Overcooked: a planning model builds a live understanding of the level while Jev agents make decisions in real time.

A live site where Laya plays Tetris, Flappy and a three-lane runner in the browser, with a measured Laya-vs-Jev benchmark and a free agent skill for local decision routing.

The author ran Jev through his LLM Chess benchmark, where a model repeatedly picks from a list of legal moves rather than generating free text. Jev's win rate against the random-move baseline landed close to several reasoning models, despite Jev not being built to reason.
Browser village where nobody follows a script. Each villager watches what you do, remembers it and decides how to treat you, answering with expressions and actions rather than talk, and a side panel shows the state sent to Jev, the questions asked and the answers returned.
Pac-Man recreation steered by Jev from structured game state. Before each junction the game sends the maze, the actors and the legal directions, and Jev answers with one direction and a probability for each option, with no screenshots, no scripted bot and no hidden pathfinding.
Work-in-progress demo of Jev driving a Minecraft agent, including fleeing zombies at night, as a test of fast structured decisions in a game loop. URL taken from a fetched awesome-jev README; thread not opened.
Tetris you can play yourself or hand over to the TypeSafe Choice primitive, which picks every move. Score, lines and level are tracked while it plays.
Nothing matches those filters
Clear a filter to widen the search, or send us the thing you were looking for.