Three vendor-noted changes create migration work for AI API integrations: Google’s Gemini Omni Flash preview deprecation date has passed, Anthropic has scheduled Claude Sonnet 4.5’s API retirement for November 30, 2026, and Google’s Antigravity Agent preview change alters some tool-call and file-edit structures. These are not three confirmed outages. A deprecation announcement or scheduled date does not establish that an endpoint has stopped working for every account.
The strict two-week window is September 19–October 3, 2026. The Gemini and Claude notices fall within it. The Antigravity announcement was September 17, just outside it, but its October 5 shutdown is close enough to warrant attention.
What changed, and what needs action
| Service or interface | Identifier and date | What to check |
|---|---|---|
| Gemini Omni Flash preview | gemini-omni-flash-preview; scheduled deprecation date September 30, 2026 |
Check whether your integration still calls the preview endpoint and test the GA model gemini-omni-1.1-flash. |
| Claude Sonnet 4.5 | claude-sonnet-4-5-20250929; deprecation announced September 30, 2026; API retirement scheduled November 30, 2026 |
Plan and test a migration. The notice is not evidence that the endpoint already fails. |
| Antigravity Agent preview interface | antigravity-preview-05-2026 is replaced by antigravity-preview-09-2026; older preview scheduled to shut down October 5, 2026 |
Update the agent identifier and, for clients using local tools or parsed function calls, review parameter casing and file-edit parsing. |
Dates and identifiers above come from the vendors’ release notes: Google Gemini API release notes, Google Gemini deprecations, and Anthropic Claude Platform release notes.
Gemini Omni Flash: the preview deprecation date has passed
Google’s release notes announced Gemini Omni Flash general availability as gemini-omni-1.1-flash on August 27, 2026, and said the existing gemini-omni-flash-preview endpoint would be deprecated on September 30. That date has passed, but the notes reviewed do not establish that Google shut down the endpoint for every user on that date. Treat it as a migration risk, not proof of a universal outage.
Recommended Free Tools
#1 Best Overall
- EVOLUTION AMD RYZEN AI MAX+ 395 MINI PC - GMKtec EVO-X2 is the next evolution in AI mini PC Ryzen Strix Halo series. Thanks to AMD Simultaneous Multithreading (SMT) the core-count is effectively doubled, to 32 threads. Ryzen AI Max+ 395 has 64 MB of L3 cache and can boost up to 5.1 GHz, depending on the workload. The Ryzen AI Max+ 395 is currently rated as the "most powerful x86 APU" on the market for AI computing.
- AI NPU with XDNA 2 ARCHITECTURE - Powered by 16 “Zen 5” CPU cores, 50+ peak AI TOPS XDNA 2 NPU and a truly massive integrated GPU driven by 40 AMD RDNA 3.5 CUs, the Ryzen AI MAX+ 395 is a transformative upgrade and delivers a significant performance boost over the competition. The Ryzen AI Max+ 395 excels in consumer AI workloads like the llama.cpp-powered application: LM Studio. Shaping up to be the must-have app for client LLM workloads, LM Studio allows users to locally run the latest language model without any technical knowledge required and unleash their creativity and productivity.
- AMD RADEON 8090S iGPU GAMING PC - The AMD Radeon RX 8060S offers all 40 CUs with up to 2.9 GHz graphics clock and uses the new RDNA 3.5 architecture. The powerful iGPU is positioned between an RTX 4060 and 4070 laptop GPU and therefore enables gaming in FHD at maximum details in most demanding games. The 8060S can also utilize the full 128GB pool, which is perfect for running LLMs such as Deepseek 70B Q8, which runs comfortably on this machine.
- EIGHT CHANNEL LPDDR5X - LPDDR5X is a new ground breaking memory small form factor installed on-board. With blazing speeds up to to 8000MT/s, it runs 1.5x faster than the DDR5 SODIMMs; 90% better performance over DDR5 SODIMMs in video conferencing and photo editing; 30% better performance in productivity apps; 12% better performance in digital content workloads.
- QUAD SCREEN 8K DISPLAY SUPPORT - EVO-X2 AI Mini PC support 4-screen 4K/8K output via HDMI 2.1 (8K@60Hz), DisplayPort 1.4 (4K@60Hz), and dual USB 4 40Gbps Transfer speed (supporting PD3.0/DP1.4/DATA). Ideal for gaming, video editing, and multitasking, it provides expansive and crisp multi-display support.
Search application code, environment variables, deployment settings, and fallback configuration for the exact preview identifier. If it is in use, test the GA identifier in a non-production environment and compare the response handling your application depends on before switching traffic.
Claude Sonnet 4.5: retirement is scheduled for November 30
Anthropic’s September 30, 2026 release-note entry deprecates claude-sonnet-4-5-20250929 and schedules its retirement from the Claude API for November 30, 2026. Anthropic recommends Claude Sonnet 5.5. The November date is the stated retirement schedule; the September announcement does not mean the existing model has already stopped responding.
Rank #2
- Built for Local AI Development: AMD Ryzen AI Halo is designed for local AI development and inference, featuring 128GB unified memory and support for up to 200B parameter models to build and run intensive AI workloads locally.
- 128GB Unified Memory: Features 128GB LPDDR5x unified memory at 8000 MT/s with 256 GB/s memory bandwidth, providing a shared memory pool across the CPU, GPU, and NPU to support larger AI models.
- AMD Ryzen AI Max+ 395 Processor: Features 16 cores, 32 threads, and Zen 5 architecture, paired with AMD Radeon 8060S integrated graphics featuring 40 RDNA 3.5 compute units and an AMD XDNA 2 NPU with up to 50 TOPS.
- Linux AI Developer Platform: Purpose-built for Linux-based AI development with full AMD ROCm software support and preloaded tools, models, and workflows optimized for local AI development.
- Compact, Connected Design: Includes a 2TB M.2 SSD, 10GbE LAN, Wi-Fi 7, Bluetooth 5.4, USB-C connectivity, and HDMI 2.1b.
Locate the model ID wherever requests are assembled, including configuration and fallback paths. Evaluate the recommended replacement against the prompts, tool use, output parsing, and application behavior that matter to your integration. The release note provides the migration direction and retirement date, not a guarantee that a replacement will behave identically.
Antigravity: check how your client consumes agent output
Google’s September 17, 2026 release notes—outside the strict two-week announcement window—say antigravity-preview-09-2026 replaces and deprecates antigravity-preview-05-2026, with the older preview scheduled to shut down October 5. This is an adjacent, near-term migration rather than a change announced during September 19–October 3.
Rank #3
- EVOLUTION RYZEN AI MAX+ 395 MINI PC - GMKtec EVO-X2 is the next evolution in AI mini PC Ryzen Strix Halo series. Thanks to AMD Simultaneous Multithreading (SMT) the core-count is effectively doubled, to 32 threads. Ryzen AI Max+ 395 has 64 MB of L3 cache and can boost up to 5.1 GHz, depending on the workload. The Ryzen AI Max+ 395 is currently rated as the "most powerful x86 APU" on the market for AI computing.
- AI NPU with XDNA 2 ARCHITECTURE - Powered by 16 “Zen 5” CPU cores, 50+ peak AI TOPS XDNA 2 NPU and a truly massive integrated GPU driven by 40 AMD RDNA 3.5 CUs, the Ryzen AI MAX+ 395 is a transformative upgrade and delivers a significant performance boost over the competition. The Ryzen AI Max+ 395 excels in consumer AI workloads like the llama.cpp-powered application: LM Studio. Shaping up to be the must-have app for client LLM workloads, LM Studio allows users to locally run the latest language model without any technical knowledge required and unleash their creativity and productivity.
- AMD RADEON 8090S iGPU GAMING PC - The AMD Radeon RX 8060S offers all 40 CUs with up to 2.9 GHz graphics clock and uses the new RDNA 3.5 architecture. The powerful iGPU is positioned between an RTX 4060 and 4070 laptop GPU and therefore enables gaming in FHD at maximum details in most demanding games. The 8060S can also utilize the full 128GB pool, which is perfect for running LLMs such as Deepseek 70B Q8, which runs comfortably on this machine.
- EIGHT CHANNEL LPDDR5X - LPDDR5X is a new ground breaking memory small form factor installed on-board. With blazing speeds up to to 8000MT/s, it runs 1.5x faster than the DDR5 SODIMMs; 90% better performance over DDR5 SODIMMs in video conferencing and photo editing; 30% better performance in productivity apps; 12% better performance in digital content workloads.
- QUAD SCREEN 8K DISPLAY SUPPORT - EVO-X2 AI Mini PC support 4-screen 4K/8K output via HDMI 2.1 (8K@60Hz), DisplayPort 1.4 (4K@60Hz), and dual USB 4 40Gbps Transfer speed (supporting PD3.0/DP1.4/DATA). Ideal for gaming, video editing, and multitasking, it provides expansive and crisp multi-display support.
Clients that execute local tools or parse function calls
For local tool execution or parsed function_call steps, the documented interface changes tool-parameter casing from snake_case to PascalCase. File edits also change from full-file rewrites to line-range replacements. Update any code that reads these structures, then test both tool dispatch and edit application; code that assumes the previous keys or whole-file format may fail to interpret the new payload correctly.
Remote sandboxes that read output text
Google’s notes say remote sandboxes that read only output_text or model_output need only change the agent identifier. Do not apply the local tool and file-edit parsing changes to a client that does not consume those structures.
How to check integrations before changing production
- Find exact identifiers. Search source code, configuration, environment variables, fallback lists, and deployment settings for the model or agent IDs in the table.
- Identify the contract your code uses. For Antigravity, determine whether the client invokes local tools or parses
function_callsteps, applies file edits, or only reads output text. - Test the replacement in staging. Exercise the actual request and response paths, including tool calls, parsing, retries, and rollback behavior. These checks are prudent integration work based on the documented changes, not vendor-reported test results.
- Verify dates and status with the vendor. Deprecation, an effective date, and shutdown are not interchangeable. Google’s deprecations page explains that a deprecated service is no longer supported and will be shut down in the near future; the listed date is the earliest possible retirement date, and exact timing may be communicated later. Check the live vendor documentation and your account’s behavior before assuming a service is available or unavailable.
What these notices do—and do not—show
The changes establish specific migration risks, not measured failure rates or evidence of widespread outages. The official OpenAI API changelog entries reviewed for September 29 document new Agents API computer-use support and GPT-6.1 Sol; they do not substantiate a quiet breaking change to existing API behavior. See the OpenAI API changelog. That distinction matters: a useful migration checklist should follow documented identifiers and contract changes, not assume every release note signals a break.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
