{"rewrite":{"id":"r_f3c99eb0d5f85e90af7b79e2","clusterId":"c_f2e9f8e3479ac61b37166f7e","slug":"typesafe-ai-ships-jev-1-14-0-with-image-input-and-262k-token-context","model":"deepseek-v4-1-flash","headline":"TypeSafe AI Ships Jev 1.14.0 with Image Input and 262k-Token Context","summary":"TypeSafe AI released Jev 1.14.0 on October 14, 2026, adding image input and a way to raise context length to 262k tokens per request. The decision model, which picks from preset options with a confidence score rather than generating text, had multimodal support disabled because non-text input performed worse. Long context stays off by default. The team also posted self-run comparisons showing higher intelligence, better cost efficiency, and a lower error rate than its previous model and other decision models.","whyItMatters":"TypeSafe AI is adding capability while telling users to distrust the numbers that would sell it, which is the unusual position of a company whose product is judged against benchmarks it does not believe in.","webCardHtml":"\u003cp\u003eJev 1.14.0 keeps long context off by default even though the model can handle far more, because reliability drops as context grows. Users opt in per request through a header, up to 262k tokens for both the request and the question.\u003c/p\u003e\u003cp\u003eImage input arrives configured through a search-and-replace function. TypeSafe AI\u0026#39;s own caveat is blunt: every model, Jev included, does worse on multimodal data than on text, and breaking questions into the simplest possible elements helps most.\u003c/p\u003e\u003cp\u003eThe release also adds box, a preview type for locating an object in an image by its four sides, and an opt-in feature for contributors to surface model inconsistencies using queries.\u003c/p\u003e","blueskyPost":"Jev 1.14.0 ships image input and a 262k-token context ceiling, both optional. TypeSafe AI acknowledges its decision model scored worse on non-text input, so the upgrade is available rather than trusted by default.","twitterPost":"TypeSafe AI ships Jev 1.14.0 with image input and long context off by default, after its own tests showed non-text input hurt the decision model.","threadsPost":"Jev 1.14.0 adds image input and a 262k-token context option, and both stay off by default. TypeSafe AI's own note that non-text input performed worse for its decision model is the part worth watching, since the features ship as user-enabled rather than trusted defaults.","newsletterBlurb":"TypeSafe AI released Jev 1.14.0 on October 14, adding image input and a way to expand context per request up to 262k tokens, with long context still off by default. The company ran its own comparisons against the previous model and other decision models, then told readers to focus on relative rankings among open models rather than the numbers themselves. The release also adds box, a preview format for locating objects in images.","attributionJson":"[{\"source\":\"GIGAZINE\",\"url\":\"https://gigazine.net/news/20261015-jev-1-14/\",\"title\":\"Jev adds image input support, long context support, and announces benchmark tests for decision models\"}]","lintFlagsJson":null,"lintHits":0,"costUsd":0,"inputTokens":4659,"outputTokens":638,"status":"published","repairAttempts":0,"nextRepairAt":null,"factsAttemptedAt":1792039956,"createdAt":"2026-10-15T04:48:57.000Z","publishedAt":"2026-10-15T04:51:10.000Z","updatedAt":"2026-10-15T04:51:10.000Z"},"cluster":{"id":"c_f2e9f8e3479ac61b37166f7e","canonicalTitle":"Jevが画像入力に対応＆長大コンテキストに対応＆意思決定モデルのベンチマークテストも発表","representativeArticleId":"a_fe928cf20bcfa8600ba98f9d","sourceCount":1,"writtenSourceCount":1,"writeAttempts":0,"isSolo":true,"entitiesJson":"{\"anime_titles\":[],\"manga_titles\":[],\"work_titles\":[\"Jev\"],\"studios\":[],\"people\":[],\"type\":\"release\",\"domain\":\"other\",\"is_roundup\":false}","contentType":"news","status":"published","firstSeenAt":"2026-10-15T02:53:00.000Z","lastSeenAt":"2026-10-15T02:53:00.000Z","updatedAt":"2026-10-15T04:51:09.000Z"},"attribution":[{"source":"GIGAZINE","url":"https://gigazine.net/news/20261015-jev-1-14/","title":"Jevが画像入力に対応＆長大コンテキストに対応＆意思決定モデルのベンチマークテストも発表"}],"entities":{"anime_titles":[],"manga_titles":[],"work_titles":["Jev"],"studios":[],"people":[],"type":"release","domain":"other","is_roundup":false},"keyFacts":null}