What changed in AI tools: week 39, 2026
21–27 Sep 2026 · 27 updates from 18 tools. Pricing and limits first.
Pricing 4
- Claude Opus 5.5 in Kiro at 2.0×, down from Opus 5's 2.2×
Experimental support for Pro and above in us-east-1 and eu-central-1, with a 1M-token context.
Source: kiro.devCurrent Kiro plans & limits - ElevenLabs for Students: three free months plus a free year of Reader Ultra
University students aged 18+ in the US, Canada, EU27, Australia and UK get free access to ElevenCreative, ElevenAgents and the API.
Source: elevenlabs.ioCurrent ElevenLabs plans & limits - Lovable chat is now free
Anyone can chat with Lovable for free to weigh ideas, plan from feedback or work on a launched app.
Source: lovable.devCurrent Lovable plans & limits - GPT-6 brings better prompt caching to cut API latency and cost
OpenAI cites higher cache hit rates, new diagnostics, explicit breakpoints and new controls.
Source: openai.comCurrent ChatGPT plans & limits
Launch 3
- Manus 2.0: new Cascade harness, Cloud Computers, Automations and Studio
Manus says Cascade used 23% fewer tokens and cost 32% less in one tested configuration; Automations can trigger on events in connected services.
Source: manus.im - Cursor launches Rollouts and Security Review bots for Teams and Enterprise
Rollouts monitors each change as it deploys and reports health per environment, and Security Review reports exploitable bugs on every pull request.
Source: cursor.com - ElevenLabs Studio 4.0 generates video, voice, music and SFX in one editor
Edit, caption and export without leaving the project.
Source: elevenlabs.io
Model 10
- Claude Sonnet 5.5 launches as the faster, cheaper partner to Opus 5.5
Sonnet 5.5 is the second model in the Claude 5.5 family, positioned as a faster, lower-cost complement to Opus 5.5.
Source: support.claude.com - Claude Opus 5.5 becomes the default for Amp's medium mode
Opus 5.5 replaces GPT-5.6 Sol in medium mode; Amp says it solved 65% of internal eval tasks for 10% less than Sol, while ChatGPT Only users stay pinned to Sol.
Source: ampcode.com - Perplexity's APIs retire older OpenAI model IDs on 24 Oct
The Agent and Orchestrator APIs drop gpt-5.4 variants, gpt-5.2, gpt-5.1, gpt-5 and gpt-5-mini at 00:00 UTC.
Source: docs.perplexity.ai - Gemini 3.8 Live gets Live Avatar for near real-time visual presence
Google adds a near real-time visual avatar to its conversational Live model.
Source: blog.google - Google releases Gemini 3.8 Flash TTS and Flash-Lite TTS
Google calls them its most expressive audio models yet.
Source: blog.google - OpenAI introduces GPT-6 Sol and GPT-6 Luna
Two GPT-6 models for everyday work that trade capability against cost differently.
Source: openai.com - Opus 5.5 in Lovable matches Opus 5 in a third to half fewer steps
Lovable's benchmarks show the same build quality as Opus 5 with fewer steps.
Source: lovable.dev - Scribe v2 Medical speech-to-text is generally available on the API
The model is fine-tuned for clinical audio.
Source: elevenlabs.io - Claude Opus 5.5 matches Fable 5.1 on most work at 40% less than Opus 5
The first Claude 5.5 model performs at Fable 5.1's level on most work and costs 40% less to run than Opus 5, per Anthropic.
Source: support.claude.com - Jasper Agent Studio adds GPT-6 Sol, GPT-6 Luna and Claude Opus 5.5
Each model is grounded in company knowledge and brand guardrails through Jasper IQ.
Source: jasper.ai
Feature 6
- Grok Team Bots: shared Grok Bots that work from a team's context
Each Team Bot combines shared context, plugins, credentials and memories, keeps each person's chats private, and can join Slack channels.
Source: x.ai - ElevenLabs opens image, video and Flows template generation to the API
Templates built in Flows can now run via the ElevenLabs API.
Source: elevenlabs.io - Developers can now submit plugins to the Claude directory
A new developer portal handles submissions and review tracking, and shows usage analytics once a plugin is live.
Source: support.claude.com - Gemini adds Connected Apps including Adobe, Airtable, Linear and Peloton
A new wave of Connected Apps is rolling out to the Gemini app.
Source: blog.google - Runway arrives in DaVinci Resolve; Kling models join Runway MCP
Paid plans get a DaVinci Resolve plugin and four Kling video models in Runway MCP.
Source: runway.com - v0's model picker opens AI Gateway models, with Enterprise owner controls
Users can browse models, review token pricing and set Thinking or Fast Mode; Enterprise teams must opt in to custom models.
Source: v0.app
Releases 1
- Midjourney alpha: live style previews, default parameters, sharper inpainting
Fast models preview your prompt across styles, defaults can be saved, and edit-model inpainting now changes only the selected pixels.
Source: updates.midjourney.com
Incident 2
- GitHub incident in Azure Central US lasts nearly 21 hours
A failed database maintenance caused API errors from 12:57 am PDT on 23 Sep, with Projects backlogs fully cleared at 9:55 pm PDT.
Source: githubstatus.com - Datacenter issue knocks out the Grok API, grok.com and Grok Build
xAI reported outages across the API, grok.com and Grok Build from about 6:03 to 6:28 pm PDT on 21 Sep.
Source: status.x.ai
Other 1
- Sora 2 and Sora 2 Pro removed from ElevenLabs as OpenAI ends the Sora API
Both models stopped generating on 24 Sep; flows using a Sora node need another video model, and Seedance 1.5 Pro retires on 11 Nov.
Source: elevenlabs.io