What people are shipping with Jev.

https://github.com/vinilana/live-jev

https://github.com/robipop22/Jev-is-odd

https://github.com/trufyrelabs/tru-jev-harness

https://github.com/raihankhan-rk/diffjury

https://github.com/AnotiaWang/awesome-jev

https://github.com/aoprisan/jev-ts-repl

https://github.com/Dujaydis/JevSysUno

https://github.com/daftAI2026/awesome-jev

@rafalwilinski @typesafeai Why it was not possible before Jev model? i am curious

@typesafeai @CompleteSkeptic @dabit3 @rileybrown @jackcheng 9) Browser use Browser Use plus Jev is stupid fast. Flight search finished in 7s for $0.0039. Fresh action space each s

@typesafeai @CompleteSkeptic @tamarajtran @0xkaushik_k @nidhisinghattri @dabit3 @Sxfyhvx @mtropolis_chris @dify_ai @MemorysaverMFA @Jasper_Black wired Coinbase data into a Kalshi B

Jev for image classification will probably break the internet and will get us to AGI @typesafeai @CompleteSkeptic

@typesafeai @CompleteSkeptic @tamarajtran @0xkaushik_k @nidhisinghattri @dabit3 @Sxfyhvx @mtropolis_chris @dify_ai @MemorysaverMFA @Jasper_Black @CompleteSkeptic on extraction casc

@typesafeai @CompleteSkeptic @dabit3 @rileybrown @jackcheng 10) Instant compaction Compaction should not be a summary prompt in 2026. Jev rates every tool call and drops the dead

@typesafeai @CompleteSkeptic @dabit3 @rileybrown @jackcheng @tamarajtran P.S. I will test Jev hard all week. I will share all learnings here. Build with these models on brewkeg and

.@typesafeai Jev 試しました まだ調整も評価も試行錯誤中ですが、PAC-MANをリアルタイムプレイしてます requestAnimationFrameベースのゲーム実装に対してgolang serverがdecision streamを送信 盤面情報は文字でstateに プレイでJevがグリッチがないようにして 利用済みtokens

Jev is absolutely astonishing - its able to classify prompts instantly and so cheap there was no point in charging users for it Huge game changer for AI from @typesafeai https://t

Jev vs GPT-5.6 Sol and Claude Haiku 4.5 at Pong. The ball moves when the model decides. Jev (@typesafeai, via @vercel AI Gateway) answers in about 200ms and makes generalised real

Jev (@typesafeai) plays pokemon, the live is here 👉 https://t.co/vans4oSIYG

这两天 AI 圈有个新模型 Jev，我觉得值得单独看一下。 它最反直觉的地方，不是 benchmark 又高了多少，而是它根本不想做一个“更会聊天”的 LLM。 @typesafeai 的思路是： 很多软件任务根本不需要模型“生成一段话”。 现在大量 AI workflow 实际上是： state → LLM → 生成文字 / JSON → par