{"schemaVersion":1,"editionId":"2026-09-23","title":"Opus 5.5 pricing, Premiere on Android, and new tools for agents and robotics","dek":"Six updates cover model costs, mobile creation, agent restrictions, open robotics, clinical transcription and evaluation transparency.","coverageNote":"Partial research for September 23, Europe/Berlin. With no completed prior cutoff, we assumed September 22 at 13:28 CEST to September 23 at 13:28 CEST. Selected announcements carry September 22 dates without times, so exact rolling-window inclusion is uncertain. X originals and parts of the broader scan remain unverified; this is not a complete account of the period.","summaryIntro":"Today's changes are practical: lower listed model prices, Android editing, and new infrastructure for longer agent tasks and industrial robots. Availability comes with limits, including a preview that cannot run guardrails and specific enterprise conditions for clinical transcription.","publishedAt":"2026-09-23T11:41:41.615Z","updatedAt":"2026-09-23T11:41:41.615Z","version":1,"items":[{"id":"claude-opus-5-5","kind":"quick-update","thumbnail":{"src":"/api/media/thumbnails/2026-09-23/claude-opus-5-5/98e7c1194c939f51a54cec7353f21e7d1e4b753f0de2220b0f9dee10f68438e4","alt":"Paper laptop and layered cards beneath the headline Opus 5.5 Arrives.","credit":"AI Daywire · AI-assisted editorial illustration","kind":"editorial-illustration"},"title":"Anthropic releases Opus 5.5 with lower listed token prices","category":"Models & Agents","tags":["models","coding"],"eventDate":"2026-09-22","status":["available","company-claim"],"sourceIds":["S1"],"short":"Anthropic lists Opus 5.5 at $4 per million input tokens and $20 per million output tokens. Its claimed 40% typical-workload saving comes from company tests.","claims":[{"text":"Listed standard input/output prices are $4/$20 per million tokens; the 40% workload saving is Anthropic's test result.","sourceIds":["S1"]}],"long":[{"text":"Anthropic's September 22 release lists input at $4, output at $20 and cache reads at $0.20 per million tokens. It says Opus 5.5 is available across its platforms, including AWS, Google Cloud and Microsoft Azure; developers can request claude-opus-5-5. Thinking mode cannot be disabled.","sourceIds":["S1"]},{"text":"Anthropic reports a 40% saving against Opus 5 on typical workloads at default settings in its own tests. That is not a guaranteed reduction in every customer's bill. Sonnet 5.5 and Haiku 5.5 are promised for the coming weeks, rather than released in this announcement. We have not tested the model.","sourceIds":["S1"]}]},{"id":"premiere-android","kind":"quick-update","thumbnail":{"src":"/api/media/thumbnails/2026-09-23/premiere-android/25d918e5e64dd397f94a72a0559f4e40f5ece699148f7552ae17bdc9f05664a2","alt":"Conceptual paper smartphone editing timeline and video frames headed Edit on Android.","credit":"AI Daywire · AI-assisted editorial illustration","kind":"editorial-illustration"},"title":"Adobe brings Premiere mobile and Firefly tools to Android","category":"Creativity & Media","tags":["video","audio"],"eventDate":"2026-09-22","status":["available","company-claim"],"sourceIds":["S2"],"short":"Adobe says Premiere mobile is available on Android 13 or later with at least 5 GB RAM. Editing essentials are free, with upgrades for extra generative credits and storage.","claims":[{"text":"Adobe announces Android availability with Android 13+ and 5 GB minimum RAM, plus paid upgrades for additional credits/storage.","sourceIds":["S2"]}],"long":[{"text":"The September 22 announcement lists Firefly-powered Generative Fill, Image-to-Video and Generate Sound Effects alongside mobile editing tools. This is Android availability, distinct from the desktop timeline update covered earlier this month. Adobe specifies Android 13 or later and at least 5 GB RAM.","sourceIds":["S2"]},{"text":"Adobe still lists continuing an edit on Premiere desktop among coming-soon features. Free editing should not be read as unlimited AI generation. We have not tested the app or its output.","sourceIds":["S2"]}]},{"id":"uipath-advanced-preview","kind":"quick-update","thumbnail":{"src":"/api/media/thumbnails/2026-09-23/uipath-advanced-preview/62eab50a992b0685363eb87f4e7f0d35c03527f9b1259364e4d5214742db1ce6","alt":"Paper folders connected by workflow arrows and a check card, headed Agents in Preview.","credit":"AI Daywire · AI-assisted editorial illustration","kind":"editorial-illustration"},"title":"UiPath previews longer-running agents, with guardrail and migration limits","category":"Models & Agents","tags":["productivity","coding"],"eventDate":"2026-09-22","status":["available","company-claim"],"sourceIds":["S3","S4"],"short":"UiPath's advanced agents preview adds a file workspace and sub-agent delegation in Automation Cloud. Guardrails are unsupported, and switching an existing agent discards some configuration permanently.","claims":[{"text":"The preview is Automation Cloud-only; guardrails are unsupported and switching permanently removes tool Static/Argument values and per-model settings.","sourceIds":["S4"]}],"long":[{"text":"UiPath's September 22 release notes introduce the advanced execution framework for document-heavy, multistep work. Its documentation limits the preview to Automation Cloud, excluding Automation Suite.","sourceIds":["S3","S4"]},{"text":"Before switching, record existing tool Static and Argument values and per-model settings: UiPath says the switch deletes them and switching back will not restore them. Agents with guardrails configured fail at startup. Delegation uses a built-in general-purpose sub-agent; custom sub-agents are not supported yet. We have not tested the preview.","sourceIds":["S4"]}]},{"id":"intrinsic-core","kind":"quick-update","thumbnail":{"src":"/api/media/thumbnails/2026-09-23/intrinsic-core/b1a545af24efda173a769d3b6fd43786267de2032a2f708bbc784880b587d9fc","alt":"Conceptual cardboard industrial gripper, blocks and engineering plans, headed Open Robotics Tools.","credit":"AI Daywire · AI-assisted editorial illustration","kind":"editorial-illustration"},"title":"Intrinsic releases an open robotics runtime and machine-tending reference","category":"Business & Infrastructure","tags":["coding","research"],"eventDate":"2026-09-22","status":["available","company-claim"],"sourceIds":["S5","S6"],"short":"Intrinsic has released Core under Apache 2.0, with ROS-compatible robotics components and an Open Machine Tending Solution reference. Robotics expertise is still required.","claims":[{"text":"The public Intrinsic Core repository documents Apache 2.0 licensing, a local runtime and ROS 2 prerequisites.","sourceIds":["S6"]}],"long":[{"text":"The September 22 release packages a local runtime, SDK and real-time control framework. The public repository also describes planning, perception and simulation components, giving developers code to inspect and adapt.","sourceIds":["S5","S6"]},{"text":"Its README specifies Ubuntu and ROS 2 Lyrical Luth prerequisites and says this is not an officially supported Google product. The reference design is a starting point, not evidence that a particular factory deployment is ready. We have not installed or tested it.","sourceIds":["S6"]}]},{"id":"scribe-v2-medical","kind":"quick-update","thumbnail":{"src":"/api/media/thumbnails/2026-09-23/scribe-v2-medical/b9dd77e34d0077846d5fa4445805b345d8dacc8b336dc0cd9c02864bea36104e","alt":"Paper microphone, transcription cards and stethoscope, headed Medical Speech to Text.","credit":"AI Daywire · AI-assisted editorial illustration","kind":"editorial-illustration"},"title":"ElevenLabs makes Scribe v2 Medical available through its batch API","category":"Tools & Applications","tags":["audio","productivity"],"eventDate":"2026-09-22","status":["available","company-claim"],"sourceIds":["S7"],"short":"ElevenLabs says Scribe v2 Medical is generally available through its batch Speech to Text API. HIPAA eligibility requires enterprise access, a Business Associate Agreement and Zero Retention Mode.","claims":[{"text":"The model uses the batch Speech to Text API; HIPAA eligibility requires enterprise customers to have a BAA and Zero Retention Mode enabled.","sourceIds":["S7"]}],"long":[{"text":"Developers select scribe_v2_medical through the batch Speech to Text API. ElevenLabs says HIPAA eligibility requires enterprise access, a Business Associate Agreement and Zero Retention Mode. General availability alone does not satisfy those conditions for using protected health information.","sourceIds":["S7"]},{"text":"ElevenLabs reports clinical benchmark gains from its own evaluations. Those vendor results do not establish accuracy in every clinical workflow. We have not tested transcription accuracy.","sourceIds":["S7"]}]},{"id":"aisi-evaluation-cards","kind":"quick-update","thumbnail":{"src":"/api/media/thumbnails/2026-09-23/aisi-evaluation-cards/d0c7c6d2db98319b2ff30165a115193267f0ff93e77fc505e5770c8626d85373","alt":"Conceptual evaluation cards and magnifying glass, headed Show Your Evals.","credit":"AI Daywire · AI-assisted editorial illustration","kind":"editorial-illustration"},"title":"AISI and EvalEval publish benchmark results with setup context","category":"Science & Research","tags":["research"],"eventDate":"2026-09-22","status":["available","company-claim"],"sourceIds":["S8"],"short":"AISI and EvalEval say they are sharing evaluation results and configuration context through Evaluation Cards. The release helps readers inspect how tests were run before comparing scores.","claims":[{"text":"The announcement describes five main benchmarks and six models, plus separate cyber evaluations with a partially overlapping model set.","sourceIds":["S8"]}],"long":[{"text":"Their September 22 announcement describes results for five main benchmarks across six frontier models, plus two cyber evaluations with a different, partly overlapping model set. The infrastructure uses the Every Eval Ever reporting schema.","sourceIds":["S8"]},{"text":"This is a transparency release, not our independent reproduction of the experiments. Its practical value is access to setup information: apparently similar benchmark scores can come from different evaluation protocols.","sourceIds":["S8"]}]}],"sources":[{"id":"S1","title":"Claude Opus 5.5","publisher":"Anthropic","url":"https://www.anthropic.com/claude-opus-5-5","publishedDate":"2026-09-22","kind":"vendor-announcement"},{"id":"S2","title":"Adobe Premiere expands to Android","publisher":"Adobe","url":"https://blog.adobe.com/en/publish/2026/09/22/adobe-premiere-expands-android-fast-powerful-easy-mobile-video-editing","publishedDate":"2026-09-22","kind":"vendor-announcement"},{"id":"S3","title":"Agents — September 2026 release notes","publisher":"UiPath","url":"https://docs.uipath.com/agents/automation-cloud/latest/release-notes/september-2026","publishedDate":"2026-09-22","kind":"vendor-release-notes"},{"id":"S4","title":"Advanced agents (Preview)","publisher":"UiPath","url":"https://docs.uipath.com/agents/automation-cloud/latest/user-guide/advanced-agents","publishedDate":null,"kind":"official-documentation"},{"id":"S5","title":"Introducing Intrinsic Core: An open source approach to Physical AI","publisher":"Intrinsic","url":"https://www.intrinsic.ai/blog/posts/introducing-intrinsic-core","publishedDate":"2026-09-22","kind":"vendor-announcement"},{"id":"S6","title":"Intrinsic Core — repository and README","publisher":"Intrinsic","url":"https://github.com/intrinsic-ai/intrinsic-core","publishedDate":null,"kind":"maintainer-documentation"},{"id":"S7","title":"Scribe v2 Medical is now available to everyone","publisher":"ElevenLabs","url":"https://elevenlabs.io/blog/scribe-v2-medical-generally-available","publishedDate":"2026-09-22","kind":"vendor-announcement"},{"id":"S8","title":"How UK AISI and EvalEval Are Making Benchmark Results Reproducible","publisher":"EvalEval and UK AISI","url":"https://huggingface.co/blog/evaleval-aisi","publishedDate":"2026-09-22","kind":"first-party-research-announcement"}],"review":{"passed":true,"reviewedAt":"2026-09-23T11:40:17.798Z","method":"source-check"},"corrections":[],"canonicalUrl":"https://aidaywire.com/briefings/2026-09-23","language":"en","aiGenerated":true,"summary":"Today's changes are practical: lower listed model prices, Android editing, and new infrastructure for longer agent tasks and industrial robots. Availability comes with limits, including a preview that cannot run guardrails and specific enterprise conditions for clinical transcription.\n\nAnthropic lists Opus 5.5 at $4 per million input tokens and $20 per million output tokens. Its claimed 40% typical-workload saving comes from company tests.\n\nAdobe says Premiere mobile is available on Android 13 or later with at least 5 GB RAM. Editing essentials are free, with upgrades for extra generative credits and storage.\n\nUiPath's advanced agents preview adds a file workspace and sub-agent delegation in Automation Cloud. Guardrails are unsupported, and switching an existing agent discards some configuration permanently.\n\nIntrinsic has released Core under Apache 2.0, with ROS-compatible robotics components and an Open Machine Tending Solution reference. Robotics expertise is still required.\n\nElevenLabs says Scribe v2 Medical is generally available through its batch Speech to Text API. HIPAA eligibility requires enterprise access, a Business Associate Agreement and Zero Retention Mode.\n\nAISI and EvalEval say they are sharing evaluation results and configuration context through Evaluation Cards. The release helps readers inspect how tests were run before comparing scores.","fullText":"Anthropic releases Opus 5.5 with lower listed token prices\n\nAnthropic's September 22 release lists input at $4, output at $20 and cache reads at $0.20 per million tokens. It says Opus 5.5 is available across its platforms, including AWS, Google Cloud and Microsoft Azure; developers can request claude-opus-5-5. Thinking mode cannot be disabled.\n\nAnthropic reports a 40% saving against Opus 5 on typical workloads at default settings in its own tests. That is not a guaranteed reduction in every customer's bill. Sonnet 5.5 and Haiku 5.5 are promised for the coming weeks, rather than released in this announcement. We have not tested the model.\n\nAdobe brings Premiere mobile and Firefly tools to Android\n\nThe September 22 announcement lists Firefly-powered Generative Fill, Image-to-Video and Generate Sound Effects alongside mobile editing tools. This is Android availability, distinct from the desktop timeline update covered earlier this month. Adobe specifies Android 13 or later and at least 5 GB RAM.\n\nAdobe still lists continuing an edit on Premiere desktop among coming-soon features. Free editing should not be read as unlimited AI generation. We have not tested the app or its output.\n\nUiPath previews longer-running agents, with guardrail and migration limits\n\nUiPath's September 22 release notes introduce the advanced execution framework for document-heavy, multistep work. Its documentation limits the preview to Automation Cloud, excluding Automation Suite.\n\nBefore switching, record existing tool Static and Argument values and per-model settings: UiPath says the switch deletes them and switching back will not restore them. Agents with guardrails configured fail at startup. Delegation uses a built-in general-purpose sub-agent; custom sub-agents are not supported yet. We have not tested the preview.\n\nIntrinsic releases an open robotics runtime and machine-tending reference\n\nThe September 22 release packages a local runtime, SDK and real-time control framework. The public repository also describes planning, perception and simulation components, giving developers code to inspect and adapt.\n\nIts README specifies Ubuntu and ROS 2 Lyrical Luth prerequisites and says this is not an officially supported Google product. The reference design is a starting point, not evidence that a particular factory deployment is ready. We have not installed or tested it.\n\nElevenLabs makes Scribe v2 Medical available through its batch API\n\nDevelopers select scribe_v2_medical through the batch Speech to Text API. ElevenLabs says HIPAA eligibility requires enterprise access, a Business Associate Agreement and Zero Retention Mode. General availability alone does not satisfy those conditions for using protected health information.\n\nElevenLabs reports clinical benchmark gains from its own evaluations. Those vendor results do not establish accuracy in every clinical workflow. We have not tested transcription accuracy.\n\nAISI and EvalEval publish benchmark results with setup context\n\nTheir September 22 announcement describes results for five main benchmarks across six frontier models, plus two cyber evaluations with a different, partly overlapping model set. The infrastructure uses the Every Eval Ever reporting schema.\n\nThis is a transparency release, not our independent reproduction of the experiments. Its practical value is access to setup information: apparently similar benchmark scores can come from different evaluation protocols.","checksum":"cca01f06e5fba02adcb3450f6a2387d228697628824859f07a61d3d7fe77c25f"}