[AINews] TypeSafe/Jev at >$100M ARR, $7.5B valuation 3 weeks after launch
https://www.latent.space/p/ainews-typesafejev-at-100m-arr-75b📌 「Decision Model」一天內被五家公司同時推出，Jev成了業界的基準線
TL;DR：Jev掀起的「決策模型」賽道一天內湧入多家對手，TypeSafe則傳出首週就破億美元ARR。
同一天內，OpenAI、Microsoft、Perplexity、Cloudflare、Liquid五家公司不約而同推出同一種東西：不輸出長篇文字，只回傳機率、選項或分數的「決策模型(decision model)」。而這整個賽道的參考基準，正是TypeSafe推出的Jev——根據Latent Space的AINews彙整，TypeSafe對外宣布了「Series AI」募資，Sequoia方面則「流出」消息稱Jev首週營收就衝破年化1億美元(ARR)。雖然也有人質疑這是否涉及炒作(astroturfing)，但這波「決策模型」的產品型態擴散速度，確實值得工程師關注。
🤔 為什麼忽然大家都在做「決策模型」
這類模型的共同特徵，是把本來要靠LLM生成長文字再解析的任務，改成單次前向傳播直接回傳型別化答案：判斷某個條件成立的機率、從清單中挑一個選項、或依等級打分。好處很直接——Agent流程中很多步驟本質上就是yes/no的判斷，不需要生成，只需要決策。
📊 一天內湧入的對手名單
| 產品 | 定位 / 特色 |
|---|---|
| OpenAI Decisions API | 三種請求型態（機率、選清單、評分），支援文字+圖片，跑在GPT-6 Luna上，$0.10/M輸入token、輸出免費，官方宣稱「快達10倍」 |
| Microsoft-Decision-1 | 定位用於LLM判官與科學假設篩選，但早期評測者指出在一致性與複雜決策上仍有困難 |
| Perplexity pplx-decider-v1.1-27b | 宣稱在1,071個案例的Decision Bench中準確率達94.5%，成本每千次決策$0.017 |
| Cloudflare clef | clef-omni支援音訊、影片、圖片與文字；clef-flash比Jev更便宜，整體速度約快2倍；權重已上架Hugging Face |
| Liquid d1 | 上架Vercel AI Gateway，支援視覺輸入的分類、路由、評分任務 |
此外，vLLM Semantic Router推出Decision 2.0，能在一次前向傳播中針對同一輸入回答多個問題並附上每個選項的機率；LangSmith則直接拿Jev當judge，針對每一筆trace分別回傳難度與正確性的型別化答案。想自己訓練的話，Unsloth釋出免費notebook，能在8GB VRAM上把Qwen3.5-4B訓練成決策模型；另一篇walkthrough指出，用Qwen3.5-0.8B在60步、4GB VRAM、約10分鐘內，準確率就能從37%提升到65%。
💡 省錢的關鍵：把「該不該生成」也外包給小模型
LangChain方面指出，把每個任務路由給「夠用就好」的最便宜模型，讓Open SWE的任務中位數成本下降了64%。這也呼應了Apple與CMU提出的Selection-based Structured Reasoning（SSR）研究方向：在agent內部，用六種自然語言策略、以長度正規化的log-likelihood在一次共享KV cache的batched前向傳播中打分。結果是單輪推理延遲下降超過90%，但端到端延遲只下降28%到54%；在Qwen3-VL-4B搭配GRPO的設定下，平均成功率61.37%，相較TAPO+GSPO基線的61.25%幾乎持平。
⚠️ 熱度與雜音並存
Microsoft-Decision-1的早期評測者已點出其在一致性與複雜決策上的弱點，顯示「決策模型」距離成熟仍有落差。而TypeSafe/Jev這則「首週破億ARR」的消息，本身也伴隨外界對炒作的質疑，Latent Space在報導中也特別提及這點，讀者應把這類募資/估值消息當作尚待驗證的業界傳聞來看待，而非定論。
🎯 實務啟示
如果你的agent流程裡有大量yes/no、分類、打分的步驟，與其每次都呼叫完整的生成式LLM，評估改用decision model（或至少把這類步驟拆出來路由到更便宜的模型/端點）可能是立即能省下成本與延遲的切入點，LangChain的64%成本下降數字就是一個現成的參考案例。
🔗 來源
- 標題：[AINews] TypeSafe/Jev at >$100M ARR, $7.5B valuation 3 weeks after launch
- 作者／機構：Latent Space
- 連結：https://www.latent.space/p/ainews-typesafejev-at-100m-arr-75b
#DecisionModels #Jev #TypeSafe #LLMRouting #AgentEngineering #AIInfra #OpenAI #Perplexity #Cloudflare #LangChain
原始資料 Latent Space · 收集於 2026-10-10
摘要原文
As you can see in the AINews X recap section below, everyone on earth has cloned the Jev API, but only one company can ever create the category. TypeSafe announced their “ Series AI ” and Sequoia “ leaked ” that they crossed 100M ARR in their first week. Although there are cynics and accusations of astroturfing , we hope it is evident that our Jev pod was 100% authentic. AI News for 10/8/2026-10/9/2026. We checked 12 subreddits, 544 Twitters and no further Discords. AINews’ website lets you search all past issues. As a reminder, AINews is now a section of Latent Space . You can opt in/out of email frequencies! Decision Models Become a Product Category The pattern : Several vendors shipped “decision” models on the same day. These return typed answers (probabilities, picks from a list, scores) in a single forward pass instead of free text. Jev is the reference point everyone benchmarks against, and @scaling01 remarked on how fast the format spread. OpenAI Decisions API : Three request types: probability that a condition is true, pick from a list, or score against levels. It accepts text and images, runs on GPT-6 Luna, costs $0.10/M input tokens with no output charge, and is “up to 10x faster” by OpenAI’s own figure ( @LearnOpenCV ). Microsoft-Decision-1 : Positioned for LLM judges and screening scientific hypotheses. An early evaluator says decision models still struggle on consistency and complex decisions ( @omarsar0 ). Perplexity pplx-decider-v1.1-27b : Claims top Decision Bench accuracy at 94.5% across 1,071 cases, at $0.017 per 1K decisions ( @perplexitydevs ). Cloudflare clef : New clef-omni accepts audio, video, image and text. clef-flash is now cheaper than Jev, and clef overall is about 2x faster ( @michellechen ). Weights are on Hugging Face . Liquid d1 : Now on Vercel AI Gateway, with vision support for classify, route and score tasks ( @vercel_dev ). Serving and routing : vLLM Semantic Router’s Decision 2.0 answers multiple questions about one input in one pass, with per-option probabilities ( @vllm_project ). LangSmith uses Jev as a judge that returns separate typed answers for difficulty and correctness on every trace ( @hwchase17 ). Train your own : Unsloth released a free notebook that turns Qwen3.5-4B into a decision model on 8GB of VRAM ( @UnslothAI ). A walkthrough on Qwen3.5-0.8B reports accuracy rising from 37% to 65% in 60 steps, about 10 minutes on 4GB ( @akshay_pachaar ). Why harnesses want this : Many agent steps are yes/no calls rather than generation. LangChain says routing each task to the cheapest adequate model cut median Open SWE cost per task by 64% ( @hwchase17 ). Related research : Apple/CMU’s Selection-based Structured Reasoning (SSR) applies the same idea inside agents ( @ZhihuFrontier ). Method : Six natural-language strategies are scored by length-normalized log-likelihood in one batched forward pass that shares the KV cache. Results : Per-turn reasoning latency falls by more than 90%, but end-to-end latency per question falls only 28–54%. On Qwen3-VL-4B with GRPO, average success is 61.37% versus 61.25% for a TAPO+GSPO baseline. Multi-Agent Orchestration and Coding Tools Claude Managed Agents dynamic workflows (public beta) : A lead agent writes a phased plan, fans it out to up to 1,000 agents per run, then merges the results. It is enabled with multiagent_20261001 ( @ClaudeDevs , config ). Cost warning : Anthropic advises starting with scoped tasks because token use can be high ( guidance ). Claude Code Projects : All waitlisted Pro and Max users were admitted. Each project runs tasks as parallel threads ( @ClaudeDevs ), and sessions can now run locally ( @gem_ray ). Opus 5.5 fast mode : It has rolled out, but it bills against usage credits and is not included in subscriptions ( @theo ). Do agent teams pay off? : Vals AI ran GPT-6 Sol and Opus 5.5 on Vibe Code Bench, alone and as teams ( @ValsAI ). Results : Teams cost 1.8–5.1x more. Only Sol at medium effort improved significantly, by 7.3 points. Behavior : Sol delegated in parallel along architectural lines. Opus ran sequential waves, reaching about 6.8 subagents and roughly 1,140 subagent tool calls per app at max effort, with no significant gain ( details ). Prime Agent rewrites itself in Rust : Over two weeks, a swarm of more than 2,000 agents used 10K+ sandboxes and 200B+ GLM-5.3 tokens. The result reaches usable input about 13x faster and uses 83% less startup memory ( @PrimeIntellect ). An accompanying essay argues that context limits lead inevitably to swarms ( essay ). Windows sandbox : A new mode built on Microsoft Execution Containers (MXC) gives faster setup, network enforcement and granular file controls ( @OpenAIDevs ). Composer predictions : Codex now suggests your next message, in beta for Pro users only ( announcement ). Some users criticize the Pro-only gating ( @Angaisb_ ). Reliability : There were complaints of daylong outages ( @dzhng ). Sentiment : DHH says GPT-6.1 Sol made Codex his primary tool over Claude ( @dhh ). Devin and Grok Bot : Devins can now spawn trees of managed Devins, so wall time tracks the slowest branch rather than the sum ( @devindevelopers ). Devin also accepts personal ChatGPT plans for GPT usage ( @cognition ). Separately, Grok Bot gets its own email address for sign-ups and scheduling ( @bot ). Model Releases and Independent Evals Qwen-Image-2.1-Turbo (open weights) : An accelerated checkpoint of the 7B Qwen-Image-2.1. It does 8-step 2K generation and natural-language editing, loads through Diffusers QwenImage21Pipeline , and launches alongside Pro and Turbo APIs ( @Alibaba_Qwen ). StepFun Step 5 Preview : A 600B-total, 27B-active sparse MoE with 1M context and vision ( @omarsar0 ). Results : It scores 33.89 on the Hermes Index, matching GPT-6 Luna, and is free on Nous Portal for a week ( @NousResearch ). Availability : It reached #1 on OpenRouter Trending, which measures usage, not quality ( @kimmonismus ). Open weights are due October 15. Max output was corrected to 64K tokens ( correction ). Upstage Solar Mini 4 : A 35B MoE with 3B active, 524K context and 208 tok/s. Its AAII score of 24 is the best at 3B active, within a point of Nemotron 3 Ultra. It is free in Cline ( @cline ). Gemini 4 Argon : Reported at 77.9% on DeepSWE v1.1 versus Opus 5.5’s 74.2%. It ships first to 650+ Fairwind Program defenders at $2/$10 per M tokens ( @dl_weekly ). Signals : Reasoning-effort selectors have appeared in Antigravity ( @testingcatalog ), and Logan Kilpatrick says “Argon is coming” ( @OfficialLoganK ). Unconfirmed : Business Insider reports that an internal “Carbon” checkpoint approaches Opus 5.5 on coding. Speech models : HeyGen Voice tops the Artificial Analysis Controlled Voice TTS arena with an Elo of 1,201, at $30/1M characters and 40 chars/s ( @ArtificialAnlys ). Whistle is a 16.9MB on-device STT model said to rival Whisper base ( @victormustar ). Multi-turn image editing : Artificial Analysis chained 30 consecutive edits ( @ArtificialAnlys ). Results : Ideogram 4.5 and FLUX 3 edit locally, leaving 95%+ of the image untouched on small edits. GPT Image 2.5 Sunburst re-renders most of the frame each turn, keeping only about 20% unchanged, so it drifts. Nano Banana 2.1 gradually darkens. OCR benchmarks : Roboflow’s new benchmark covers 48 models, with GPT-6 Astra leading text localization ( @skalskip92 ). Datalab’s OmniParseBench has 16K tests across 90 languages, and its own model does not rank first ( @VikParuchuri ). Arena roundup : Claude Haiku 5.5 ranks #30 on WebDev at $0.10/$0.50, matching GPT-6 Luna’s price while scoring 6 points higher. Mistral Large 4 sits at #43 on Agent Arena ( @arena ). On ARC-AGI-3, a new high score of 59.17% ( @arcprize ). Research, Training and Inference Systems vLLM and SGLang on Vera Rubin : vLLM reports more than 7.8x GB200 throughput on MiniMax M3 at matched interactivity on AgentX. These are early results ( @vllm_project ). Technique : Locality-aware MoE uses CUDA 13.
由 claude-code/sonnet 自動生成