Jev explained: where to start

The list below is one path through the digest. Every figure on those pages is a claim from the people who published it unless a third party ran the same test.

1. What it returns

What Jev returns is the explainer: Choice, Score, Noul, and the claim that the model does not write text. The launch story is Almeida's September 15 announcement and TypeSafe's own workflow table. Jaggedness is the company's list of ways 1.13 still fails. Limits keeps that list and adds misses that showed up in later stories. Decision model hallucination separates the launch post's can't-hallucinate line from a wrong option inside the schema. What Jev is puts the same description, the comparison with a language model, and the fine print on one page. Jev Day 2026 is the October 14, 2026 release: jev-1.14.0 as jev-preview, images behind TypeSafe-Danger-Zone-Images, Box on 1.14.0, and 262k context. Opened October 15, OpenRouter listed typesafe/jev-1.14. At 08:19 UTC October 16, one Box for every vehicle returned one rectangle, n = 3, and type array returned HTTP 400. Date headers 08:49:11 GMT to 08:49:40 GMT October 26, jev-latest named jev-1.14.0. At 00:28 UTC October 16, TypeSafe's jev-latest still returned jev-1.13.0. Aviz Maeir scored Jev 1.14 Box at 82% on 180 scenes. Yamashita scored 0.833 on 1.14.0 against 0.817 on 1.13.0 on 60 Kyoto-ben tickets. Jev4Mellea's 12 scenes on 1.14.0 covered 5 of 12 at 0.10 / 0.90. Decision models like Jev lists the other hosted calls and open checkpoints, including OpenAI's Decisions API and Cloudflare's Clef. OR-decision-4B prints JevBench 193 of 231, 83.55%, with launch-chart Choice 79.7 against Jev 79.6. MiniCPM5-2B-Jev prints 78.79% on JevBench, first among open models at or under 2B. BranchOne returns Choice probabilities from Qwen3.5-0.8B-Base, median 0.305 s on ten short text questions. Open decision models is why the published checkpoints start from a public language model.

2. Where to call it

Access lists the public doors this desk has already cited: TypeSafe's API, Vercel AI Gateway, Cloudflare Workers AI, Netlify, OpenRouter, Venice, and LiteLLM. Clef and Clef-flash are Cloudflare's own models on Workers AI, posted October 1, at @cf/cloudflare/clef and @cf/cloudflare/clef-flash. The typesafe/jev card on that host is a separate listing. On October 9 the Clef blog lists Clef-flash at $0.038 per million input tokens and Clef-omni at $0.15. PurpleMIST-Flash-1.0 scores 0.668 on 15,050 translated African-language decisions. vegaml scores 75.4% on 800 phishing emails against Jev 1.13.0 at 61.9%. hunch, posted October 27, 2026, turns EmbeddingGemma2 into a CPU System One endpoint at 20 ms on text and 100 ms on images, with Cloudflare Clef as a real-time teacher. Opened October 28, public search found no hunch GitHub repository. d3-ultra prints Decision Index 0.4.1 Full 70.18, second of 229. d3-turbo prints Full 67.31 and vision full 72.46. GLiDE v1.2 prints Decision Index 0.4.1 Full 71.5, rank 1 of 229. GET /v1/models omitted GLiDE. Jeva-4B prints 0 flips on 600 Choice items. mammoth-1, posted October 21, 2026, prints Decision Index 0.4 Full 71.24 on the card. Opened October 23, live 0.4.1 Full 68.4 rank 3. GET /v1/models with no key returned HTTP 401. d1-ultra, posted October 15, 2026, prints 65 on Artificial Analysis Decision Index against Jev 1.14 at 57, opened October 16. Vals AI printed all 1,960 ClaimBench claims right on October 22. Liquid prices input at $0.042 per million tokens. This desk did not call that API. Celeris-1 Decision, posted October 8, 2026, prints 80.0% on 3,210 jev-bench items at a 67 ms median, and $0.04 per million input tokens. Opened October 16, Artificial Analysis Decision Index lists it at 52. GET /v1/models with no key returned HTTP 401. Span-02 prints Decision Index 0.3 thinking off at 61.2 against Jev 1.13.0 at 58.0. The card prints text index 63.62. This desk did not run the weights. OpenRouter Armor scored Jev 1.13 at 90.7% on 492 tool calls, 446 of 492. SubQ Snap 1.1 prints 65 on Artificial Analysis Decision Index. Docs name subq-snap-1.1 at $0.04 per million input tokens. Date headers 02:23:17 GMT to 02:23:20 GMT October 23, three POSTs from this desk returned HTTP 200. Opened October 19, 2026, JevBench ranks basal-omni-1.0-ultra first of 184 open systems on Capability at 82.5, and kettei-fat first of 38 hosted systems at composite 77.8. Jev 1.14.0 is ninth at 70.2. Date headers 08:49:11 GMT to 08:49:40 GMT October 26, TypeSafe's jev-latest named jev-1.14.0. On October 19, jev-latest still returned jev-1.13.0. Docs price kettei fat at $0.06 per million tokens. GET /v1/models with no key returned HTTP 401. Drex Max, posted October 19, prints Decision Index 0.3.2 public 67.59 against Jev 1.14 at 60.70, at $0.06 per million input tokens. GET /v1/models with no key returned HTTP 401. StartLux-Decision v1.1 27B Instant prints Full score 69.12, first of 174 on Decision Index 0.3.2. Sieve-31B prints Full score 63.8. Clairic is Cloudflare's extraction model. The cascade table is 19.3 s against 61.7 s. Datadog prints Jev Noul recall 82.90% to 3.88%. Date headers 08:50:46 GMT and 08:50:47 GMT October 20, this desk's jev-1.13.0 is_risky was 0.94 then 0.13. HiddenLayer prints Jev at 91% Confirm Price on a checkout ad. Date headers 07:24:57 GMT and 07:24:58 GMT October 19, this desk's jev-1.13.0 still chose place_order, 1.0 then 0.89. ToddlerJev prints 68.70% on MMAU-mini. Qwen-Audio-Jev prints 78.70%. pplx-decider-v1.2, posted October 16, prints Decision Index 0.3 overall 68.09 on joint requests, at $0.02 per million input tokens. Opened October 11, 2026, that board ranked Sage 1.3.0 first of 27 at composite 74.0, and Microsoft-Decision-1 sixth at 69.1. H2O-Lightning-4B v1.1 prints 72.5 against Jev at 71.5 on its October 7 blog. OpenBles Firewall 1 caught 119 of 120 agent attacks. d3 prints 66.3 on Decision Index 0.3.2, opened October 16. The October 11 reading was 64.7 on edition 0.3.1. Mizan 4B averages 78.7 on 3,000 Persian records. THX-01 averages 98.4 on 2,843 support tickets. Quyet-1.0-Large prints Capability 81.7 on frozen JevBench v1.6.0, rank 1 of 92 on that page. Blink v0.3 prints 57.48 on a self-run of Decision Index 0.2.1. TypeSafe posted on September 20 that the waitlist is gone and that new accounts start with $5. On September 21 it posted an API and console incident, then said it was back; the status page marks Console is unavailable resolved at 08:16 UTC. On September 22 TypeSafe paused new signups and said existing accounts keep working. On September 24 it said the signup banner was removed and that signups were still closed. On September 25 one person posted a logged-in console, and an hour later another wrote that new registration was still stopped. On September 27 at 22:30 UTC the company posted that signups were open again, and that new signups no longer get free credits. On October 6, 2026, at 17:05 UTC, this desk signed in to a new console account. Available credits were $0.00. Price, window, and quota are whatever that host published. Jevflake is a dbt package that calls TypeSafe from Snowflake SQL. duckdb-jev is a native C++ extension that batches the same calls from DuckDB. MotherDuck's prompt_jev() is a hosted SQL function on paid plans. Vercel's form router calls Jev through AI Gateway and keeps the destination when confidence is at least 0.95. typesafe/jev-router is OpenRouter's chat model id. Arena's October 7 thread prints +3.6% net and a median $0.155 a task on more than 4,700 sessions.

3. Who measured what

GEV-26B-Decide prints an adaptive 62.48 on the authors' Decision Index 0.2.1 scoring, against the September 28 board's Jev cell of 57.91. Adaptive thinking on Knowledge misses the kit's latency cap. hunch prints 20 ms on text and 100 ms on images for an EmbeddingGemma2 System One endpoint on CPU, posted October 27, 2026. Opened October 28, public search found no hunch GitHub repository. d3-ultra prints Decision Index 0.4.1 Full 70.18, second of 229, opened October 26. d3-turbo prints Full 67.31 and vision full 72.46. GLiDE v1.2 prints Decision Index 0.4.1 Full 71.5, rank 1 of 229, opened October 26. Docs $0.04 per million input tokens. The October 1 self-run of 64.81 stays. Jeva-4B prints 0 flips on 600 Choice items. BANKING77 at K=16 is 91.62 against Jev 82.50. Decision Model Factory prints DF-JA-Large at 67.63 on S1MB Japanese Task Avg. Abe's 1,774 posts cost $0.055 on Jev Noul against $1.87 on AI_FILTER. pplx-decider-v1.2 prints Decision Index 0.3 overall 68.09 on joint requests. The October 1 panel still prints 85.71% on a 7,210-row panel, against jev-1.13.0 at 84.51%. Decision Index 0.3, opened October 7, 2026, ranks Perplexity Decider v1.1 first at 62.8, with Jev 1.13.0 as a reference at 60.1. The quickstart opened that day prices the API at $0.02 per million input tokens. Drex DLM, posted October 6 as Drex 1.1, scores Decision Index 0.2 at 52.31 against the kit's Jev cell of 51.67. PolicyLM-1.7B prints custom-policy accuracy 0.842 at a 0.5 cutoff, with no Jev column. HiringCafe's October 7 thread, on the OpenAI Jev page, prints Spearman 0.74 for Jev and 0.71 for OpenAI on query-to-job relevance. Databricks ai_decide is a Beta SQL function. The docs do not name the model and do not print a price. NIRNAY 450M prints 0.8792 on 3,080 Banking77 rows. The Jev cell was not measured in that repo. WaterSheep prints 77.8% in distribution and 61.2% on held-out datasets, with no Jev column. StartLux-Decision v1.1 27B Instant prints Full score 69.12, first of 174 on Decision Index 0.3.2. The October 3 self-run of edition 0.2.1 stays 63.88. basal-1.5 prints 72.12% on Werdykt v1, against Jev 1.13.0 at 81.64%. Decision 2.0 Vega prints 56.5 on its own Decision Index 0.2.1 run. The card does not print a Jev row. Vela 2.0's 9B card prints 41.63 on Decision Index 0.2.1 with Noul calibration on, and 41.09 with it off. That file prints Lux at 46.23 and does not print a Jev row. fsdecide scores 91.3% of 80 review items against jev-latest at 56.3%, and 60.5% against 90.5% on the authors' general set. Tev1, opened October 7, 2026, is $0.04 per million input tokens on Together's page, and the chat call returns one letter. Ollama serves it on /v1/systemone. Mercury Decide cites 14 decisions a second and prints no score. Opened October 8, 2026, the API board ranks it third at 72.4, and the free model page returned, priced Free. mammoth-1 prints Decision Index 0.4 Full 71.24 on the card. Opened October 23, live 0.4.1 Full 68.4 rank 3. d1-ultra is the only decision model Vals AI printed with all 1,960 ClaimBench claims right. Opened October 16, Artificial Analysis Decision Index prints 65 against Jev 1.14 at 57. Celeris-1 Decision prints 80.0% on 3,210 jev-bench items at a 67 ms median. That board lists it at 52. Span-02 prints thinking off at 61.2 against Jev 1.13.0 at 58.0 on Decision Index 0.3. OpenRouter Armor scored Jev 1.13 at 446 of 492. SubQ Snap 1.1 prints 65 on Artificial Analysis Decision Index, at $0.04 per million input tokens. Date headers 02:23:17 GMT to 02:23:20 GMT October 23, three POSTs from this desk returned HTTP 200. Microsoft-Decision-1, posted October 9, 2026, lists $0.042 per million input tokens and says it was 4.5 times quicker than Quyet-1.0-Large. The post does not print an accuracy. The Foundry catalog, opened October 10, says the weights are not distributed. TypeLLM scores Qwen3.8-27B at 228 of 231 public JevBench tasks with thinking. The same file has 195 of 231 without thinking, and the hard tier moves from 76 of 111 to 109 of 111. The homepage prices input at $0.05 per million tokens. That run is not the hosted name typellm-latest. JEV-27B-VL finishes 15 of 20 pick-and-place scenes. Its six-benchmark mean is 84.07 against a hosted Jev cell of 83.85, and that JevBench column is not the official board. On October 4 a FireRed post on the pplx-decider page reports 137 calls and an estimated $0.028, and it prints no Jev cell. jevmod scores sexual text at AUROC 0.984. ZeroNSFW scores pictures on the host. StreamDecisionBench prints Kev-27B at 62.15 and the Jev API at 59.63. A CUA Speedrun chart prints Jev at 9.8% on 50 OSWorld tasks, 125 seconds each. A 37-dataset abstract counts 346,009 requests for under $10.

Strands Decider 2B, posted October 1, scores 167 of 231 on the public JevBench tasks, with a median 115 ms on an RTX 3090. That rank is the September 25 v1.4.2 board. HydroJEV reports jev-1.13.0 at mean macro-F1 0.63 on simulated in-distribution alarms, and a gate that settles 69 of 200 and 76 of 200 fresh windows. A SemEval pipeline prints Task 1 RMSE 1.0645 from frozen Jev and 488 fitted coefficients. jev-doc-search marks both of two filing questions correct. The income-statement page is among the pages kept.

WrangleBench prints 137.3 perfect matches per 200 with Jev and 105.1 without. An Insilico chart on 12,563 molecules puts a gradient-boosting ensemble above Jev 1.13 on five CYP assays, and Jev's probabilities stop at 0.69. You.com's three risk sweeps cost $0.43 in all. A blind read of the reports ties at 4.25. Bud Decision Studio serves eleven local models at port 8420. DSPy 3.4.0 calls Jev from Predict. A patterns repo pins jev-1.13.0 and a shell gate at 0.2 and 0.6.

Evals collects the numbers already in the stories, with the caveat that came with each one. JevBench is a 534-decision suite that mixes accuracy with calibration, speed, and cost, so GPT-5.6 Luna can win the hard items and still sit below Jev on the composite. The v1.2.8 table is a geometric mean: Jev 75.4, SemIf 74.7, djev 74.3. v1.4.1, posted September 24, adds 308 sealed decisions and a harmonic mean. Jev is 63.29 there, with public accuracy 86.6% and sealed accuracy 36.7%. v1.4.2.2, changelog September 27, keeps those Jev cells and ranks Imajev-4B first at 67.37. v1.5.0, posted September 28, is a new board: 1,624 decisions, Cygnet 73.7, Winnow-12B Q8 73.2, Jev 72.1 at official rank 3. The page calls the top pair a statistical tie. Jev's capability on that page is 80.0, first among 56 Jev-class systems. 63.29 is not the v1.5 score. The API board opened October 7, 2026, release v1.6.1 under revision v1.7.12, ranks Jev 1.13.0 third at composite 71.5, with capability 77.1. Opened October 8, 2026, revision v1.7.21, that board ranks 26 hosted systems and Jev 1.13.0 is fourth at the same 71.5. The v1.5.6 page opened October 4 lists imajev-4b at 70.4, official rank 10. Imajev reads up to two photos. Plumb prints 89 of 111 on its own public-hard runner. A 400M encoder named Kev scores 79.96% on 17,476 of its own rows, against Jev at 73.54%. jevgrep keeps 8 of 10 SWE-bench solves and prints a 25.8% lower combined bill. toolgate 0.15 scores 19 of 20 on ten matched pairs. A grants thread prints 75 of 78 nos hidden. jevbench.app replays 20 of 21 recorded answers. That site is a different project. Decision Index 0.2 is a different average: the kit's test locks Jev at 51.67. The September 28 board file is edition 0.2.1 and lists Jev at 57.91. The page we opened September 30 is JevBench v1.5.4, with 106 ranked systems. The September 29 addenda are numbered on that page and still tagged roster addendum. Matilda prints a self-run Decision Index score of 59.26. That 59.26 is not in the September 28 board file. Liquid's d1 chart is a separate reproduction at 58.9 against Jev at 57.9. Open d1 prints d1-3B at 48.57 on a self-run of Decision Index 0.2.1. That table has no Jev row. Step-Jev scores a Qwen3-4B search agent at 19.6% when every step is graded, and 5.4% when only the final answer is. A bag of Jev questions moves the WANDS mean from 0.574 to 0.610 and leaves the median at 0.561. Drex 1.5 prints 199 of 231 against Jev at 201 of 231 on 231 public items. The September 30 post calls that a JevBench lead. Rox's chart keeps 96% of relevance at 200 chunks, against 84% for a production reranker, mean over 75 queries. Jev Sentinel is the demo behind TechCrunch's $2.94 against $372. The demo page prints about $55 per million actions. Mercury Decide is Inception's September 30 claim of first place on OpenRouter's JevBench v1.4. The post prints no score. Opened October 8, 2026, the API board ranks it third at 72.4, and the free model page returned, priced Free. OpenAI's Decisions API is a separate call. TechCrunch on September 30 quotes Sam Altman on speed and says TypeSafe did not answer its questions. OpenAI Jev records Date headers 02:23:12 GMT and 02:23:13 GMT October 23, two POSTs HTTP 200. OpenAI Developers posted general availability on October 22. The guide prices uncached input at $0.10 and cached input at $0.01. A cat sentence on that page scored 0.98 on Decisions and on Jev. TypeSafe's Deel charts put 6 of 8 quality checks at parity or better, and a 16.6-point drop on messy metric questions. A registered attack paper moves validated success on jev-1.13.0 from 1.8% to 3.5%. RLCDAlignBench puts a generic Noul at median AUROC 0.886, and median ECE at 0.168. jevos-1b prints 0.815 against Jev at 0.927 on 2,000 yes/no questions. jev-code still fails three coding tasks after the call-nesting fixes. astra-jev-harness cuts Astra input 27% on one two-file fixture. An older run cut files from 28 to 4 while Astra input fell 1.2%. CLM-8B is a dual encoder on a frozen Qwen3-8B. Its verifier holdout is 31 of 38 DeepSWE tasks. Eikos-27B prints 82.9 on the author's public hard row, against Jev at 73.0, and 2.4% error once confidence is 0.90. A September 24 chess screenshot scores 388 games at 53.9% for Eikos. mizchi's fit table counts 0 illegal moves in 489 MOBA calls, and 6 to 27 wrong pages where a one-line rule was 16/16. JevEval is DeepEval's custom metric. On October 14, 2026, at 01:58 UTC, this desk's jev-1.13.0 calls scored its Tool Faithfulness example at 0.254 when the answer invented a breeze, and 0.901 when it did not. TrueForge is Akshay Pachaar's weekly client follow-up agent. On 40 matches, Jev and Opus both scored 37 of 40. Printed decision costs are $0.0015 and $0.005 to $0.015. He wrote that 40 is too small for the confidence cut. On October 14, at 02:34 UTC, this desk's jev-1.13.0 call picked the intended issue at confidence 1.0. LangChain's judge eval puts Jev next to three LLM judges on five frozen traces. AstroHan's 30-task filter is the first keep-or-drop run on this desk that reports whether the agent still finished the task. Aman Kumar's 16,000-call write-up is the longest independent measurement we have, public cases included. Charly Poly's encoder thread is the first comparison against open zero-shot classifiers rather than a generative model. DocJev's 40-PDF pilot puts Jev next to Luna on classify and split of the same LiteParse text. RZ's 100-case security demo puts Jev next to Terra and Opus on synthetic TypeScript spans: 65.3% balanced accuracy against 75.1% and 91.2%. jevsearch's 41-query table is a ranking bench on TypeSafe's docs, 83% Hit@1 after Jev against 41% on keywords. Open-Jev's public subset scores 231 of 534 JevBench tasks: released 9B 179/231, 27B v1.1 197/231, Jev 200/231, Astra 231/231. jsort reports Pearson r 0.824 on 300 CommonLit excerpts. Pairwise Choice on eight Stripe passages this desk wrote put Roll secret at 8.86 on a 0-9 Score and left the other six at 2.42 or below. MotherDuck's AG News table is 89% on 100,000 training-split rows, 40 seconds, $0.50. jev-gateway's chess bench records a 57% drop in bug-fix output tokens for Astra and Sol, and a slower feature task for Opus 5. AgentOS's 121-turn chart puts an opt-in Jev router at accuracy 0.901 against its local classifier at 0.355. laya-mlx times the Laya checkpoints on an M3 Max: 13.42 ms and 7.39 ms, and matches upstream answers on 63 of 63 validation questions. Spring's Java client records a 310 ms three-question call from one laptop. Release 0.3.0, September 30, points the same client at Laya. The Ollama comparison is on the snapshot docs. We did not run it. Decider 1 posts accuracy 0.768 against Jev 1.13.0 at 0.727 on teacher-average labels. dejevu books the Flights search in 5.63 seconds on Groq. The Jev cell beside it, 7.09 seconds, is the other repo's published figure. A 612-row banking feed has Jev at 36.1% and Gemini 3.7 Flash at 42.5%. poker-table times a 46-spot bank at about one second median. The post's live pill says 545 ms. rev posts public hard-tier accuracy of 0.784 for a frozen Qwen3.8-27B, against a Jev cell of 0.730 in that README. The laptop 2B row is 0.550. spark-s1-4b-v6 prints a public-proxy Intelligence of 83.1 and hard-tier 0.595. hotdog-27B is 1,603/1,631 on its binary holdout and 60/74 on public binary items, where Jev is 62/74. A Japanese table has Jev at 96.0% on JCommonsenseQA and 56.0% on JamC-QA. Twenty-five probability words line up with 99 people at rank agreement 0.99. ASO Workbench reports Spearman 0.74 against Apple on 578 held-out terms, using Jev as a feature. Jevsus is a true/false diode on jev-1.13.0: the post counts 209,628 calls and $3.13, and 45% of statements move by 0.20 or more across three wordings. Rizzo Flow scores SemIf's authored144 at 0.812 against SemIf's published 0.819, at 49 ms on an RTX 5060 Ti. Jevre posts 0.9323 on its test split and 0.8366 on held-out sources, with no Jev column. Sarashina on JGLUE test is 34% and 86% against LFM at 17% and 69%. Ámbar Reflex has the 4B model at 24.9% of the level and the 9B model at 82.7% offline. One screenshot is $0.0002 on Jev and $0.032 on Opus 5. Bryo's 1,565 emails is a named-company table with a Gemini baseline. Compare puts those similar tests next to each other. The rows are not a ranking.

4. How people put it in a loop

Patterns groups the setups that show up more than once: rank a legal set, judge a pile of documents, override only when confident, route the expensive call, recreate the model locally. Browser Use flights is the early click-loop demo. Jevis is the same idea inside Flutter's test harness. SemIf is a 4B open rebuild that reads option logits and sits 0.7 behind Jev on JevBench. Laya is an Apache 2.0 ModernBERT-sized encoder; JevBench 70.1, and its own table trails published Jev on Banking77. Kev is a LoRA plus pointer head on Qwen that serves the same HTTP shape; author new-source table Kev-8B 79.6% versus Jev 85.7%. djev is a DiffusionGemma decision API with image input; JevBench 74.3. DiffusionGemma as Jev is the one-step canvas read behind that idea. A hand chart adds to 198 of 201 against Jev at 191 of 201. Clef is Cloudflare's Apache 2.0 pair on frozen Qwen3.8-27B and Qwen3.5-9B. The blog median is 209.3 ms and 38.8 ms against a printed Jev cell of 524.1 ms. SimpleJev reads next-token logits from a stock open model; JevBench Qwen3.8-27B 67.3. Open-Jev is a Qwen3.5 LoRA plus a decision head. On the 231 public tasks the 9B checkpoint is 179/231, and 27B v1.1 is 197/231 against Jev at 200. jevctl wraps Jev as shell commands, including a compact that keeps kept lines verbatim. DocJev classifies and splits PDFs with local LiteParse and hosted Jev; the author's table is 40/40 classify and 7/8 split against Luna. n8n-nodes-jev-classification is a community node with a Needs Review branch. duckdb-jev asks from SQL inside DuckDB. jevsearch re-ranks a docs search after a keyword pass. jsort orders a file from pairwise Noul judgments. Pairwise Choice asks which of two passages is more relevant, then fits Bradley-Terry. The form router hands the uncertain band to Luna Fast. llamacpp-jev serves the same HTTP shape from stock llama-server, images included, with a raw label softmax. A reply the same hour links kishida/llama.cpp branch jev, which compiles that path into llama-server. On October 2, upstream llama.cpp merged it for five decision-model GGUFs. Release v0.6.0, on October 5, adds Nimble and Clef. Clef can take an image. The branch page, opened October 6, still says the endpoint is not upstream. jym asks Jev only when a subcommand is missing from help text. jev-gateway asks Jev which tool a coding agent should call, and can skip the language model when every argument is closed. laya-mlx runs those Laya weights on Apple Silicon without a new accuracy table against Jev. Spring AI TypeSafe can refuse a request before the chat model runs, and can send a failed criterion back for another try. AgentOS asks Jev which model tier a turn deserves, and leaves the local classifier as the default. jev-rubiks lets code solve the cube. Jev decides whether the coach should speak, after a test where picking turns left the distance at 22. Exa's demo puts search snippets into Jev state and does not print the with-search and without-search pair. Mejias's pg-jev asks from PostgreSQL and returns jsonb, with no scored set in the README. Zachi's pg-jev is a different extension: about 3.5 seconds and $0.012 on 2,000 rows, then about 50 ms from cache. jevelry maps each answer to act, mark, or fall_back from thresholds in a JEVEL.md. poker-table asks Jev for a read of the opponent and keeps the bet math in code. swift-jev is a Swift client whose sample policy escalates under 0.60 and acts at 0.85. djev-run serves DiffusionGemma-Jev on Cloud Run at a stated 35 to 60 ms, with no accuracy table. postgres-search can sink a bad SQL hit and says that Jev pass was not measured. rev and spark-s1 are two more local readouts of open weights, with public-tier tables of their own. Rizzo Flow is Spark-X2.5 on llama.cpp, 26 options, and 6 of 36 missing-evidence rows still confident. Jevre is an 0.85B checkpoint on the TypeSafe Python SDK. Shuffling the state drops the test score from 0.9323 to 0.4282. jev_local adds Sarashina2.2 Vision 3B beside LFM, with a 26-candidate cap. Ámbar Reflex plays a run-and-gun from six typed questions on an RTX 3060. awlevin's loop OCRs the Mac and asks Jev for the next action. Dates are parsed in code before that call. dejevu runs the flights loop on a chat model and leaves its Jev backend untested. Jevonian asks Jev which model should take a coding turn, and prints no scored table. JevGate asks about one function at a time and composes the finding at 0.80. A 28,074-sentence panel adds Jev as a fourth judge. The odds ratios stay above 1, and the note has no gold label. A 100-item post scores Laya's English checkpoint at 92 and Jev at 100. JEV Search reranks a local folder. NFCorpus nDCG@10 moves from 0.322 to 0.377. Jev RAG puts the same kind of gate next to TypeSafe's CLERC note, top-1 5% to 18% on 40 queries, and Hindsight's LoCoMo recall@1 of 0.950 against 0.800. A translation bake-off uses one Jev pass as the ruler: 1,072 scores, $0.3165. A 100,000-line needle takes 1,081 seconds with Jev and 1,480 without, and both runs return 2867825. Five runs of 42 claims leave Jev at 73.3% and GPT-5.4 at 70.0% on the hard 18, with McNemar p = 1.000. Florian's JevBench table is the separate 534-task composite. semgate is a Go HTTP middleware with a public range attached.

5. Indexes

Timeline is the same stories by article date. Limits is the failure list. Compare is similar tests side by side. People is who published them. Repos is the GitHub projects those stories cite. Terms is the glossary. Sources is the URL list. Method is how a story gets on the desk. FAQ is short answers from the same material.