Agents API is in public beta π
OpenAI put the Agents API in public beta. Cloud agents with the Codex harness, managed sessions, tools, and either a hosted sandbox or one you bring yourself.
What shipped:
β Codex harness, not a chat-completions wrapper you have to babysit
β Managed sessions and tools in the API
β Hosted sandboxes, or bring your own
Public beta, so expect sharp edges. If you ship coding agents or anything that needs a real environment, this is the post to read. Not another model card. Call it if you are done stitching sessions, tools, and a VM by hand.
[ Read More ] :
https://openai.com/index/introducing-the-agents-api/
γ°γ°γ°γ°γ°γ°
#AI #LLM #Agents
@ProgrammingTip
OpenAI put the Agents API in public beta. Cloud agents with the Codex harness, managed sessions, tools, and either a hosted sandbox or one you bring yourself.
What shipped:
β Codex harness, not a chat-completions wrapper you have to babysit
β Managed sessions and tools in the API
β Hosted sandboxes, or bring your own
Public beta, so expect sharp edges. If you ship coding agents or anything that needs a real environment, this is the post to read. Not another model card. Call it if you are done stitching sessions, tools, and a VM by hand.
[ Read More ] :
https://openai.com/index/introducing-the-agents-api/
γ°γ°γ°γ°γ°γ°
#AI #LLM #Agents
@ProgrammingTip
OpenAI
Introducing the Agents API
Build and launch cloud agents with the Agents API, a managed service powered by the Codex harness for orchestration, long-running sessions, and tool use.
GPT-6 Astra is OpenAI's new work model π
OpenAI shipped GPT-6 Astra. They are calling it the next generation in intelligence for work.
Not a research teaser. A work model. Expect the name in agent stacks and coding tools.
Why you care:
β New flagship name to put on your eval harness
β Pitched for workplace tasks, not playground chat
β If you pin model ids in config, read this before you swap anything
Start with the official writeup. Then decide if your coding agents still belong on last week's snapshot. I would not change production on a blog title alone. Run your own traces.
[ Read More ] :
https://openai.com/index/gpt-6-astra-next-generation-work
γ°γ°γ°γ°γ°γ°
#AI #LLM #OpenAI
@ProgrammingTip
OpenAI shipped GPT-6 Astra. They are calling it the next generation in intelligence for work.
Not a research teaser. A work model. Expect the name in agent stacks and coding tools.
Why you care:
β New flagship name to put on your eval harness
β Pitched for workplace tasks, not playground chat
β If you pin model ids in config, read this before you swap anything
Start with the official writeup. Then decide if your coding agents still belong on last week's snapshot. I would not change production on a blog title alone. Run your own traces.
[ Read More ] :
https://openai.com/index/gpt-6-astra-next-generation-work
γ°γ°γ°γ°γ°γ°
#AI #LLM #OpenAI
@ProgrammingTip
OpenAI
GPT-6 Astra: The next generation in intelligence for work
Meet GPT-6 Astra, OpenAIβs most capable model for business, with advanced reasoning, computer use, and stronger writing and design judgment.
Gemini 3.8 Live is generally available π
Google moved Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking to general availability in the Gemini API.
These are the real-time audio models exposed through the
What to check:
β’ Gemini 3.8 Live: the lower-latency voice model.
β’ Extended Thinking: use it when the spoken request needs more reasoning before the reply.
β’ GA status: worth revisiting if you held off on a preview-only voice feature.
For .NET teams, this is a good fit for a streaming WebSocket service, not a request-response controller.
[ Read More ] :
https://ai.google.dev/gemini-api/docs/changelog
γ°γ°γ°γ°γ°γ°
#AI #Gemini #LLM #API
@ProgrammingTip
Google moved Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking to general availability in the Gemini API.
These are the real-time audio models exposed through the
Live API. Live is for the voice path: streaming audio in, streaming audio out, and handling a conversation without bolting together STT, an LLM call, and TTS yourself.What to check:
β’ Gemini 3.8 Live: the lower-latency voice model.
β’ Extended Thinking: use it when the spoken request needs more reasoning before the reply.
β’ GA status: worth revisiting if you held off on a preview-only voice feature.
For .NET teams, this is a good fit for a streaming WebSocket service, not a request-response controller.
[ Read More ] :
https://ai.google.dev/gemini-api/docs/changelog
γ°γ°γ°γ°γ°γ°
#AI #Gemini #LLM #API
@ProgrammingTip
Google
Release notes | Gemini API Documentation
Keep track of updates to the Gemini API
Hex turns GPT-6 Astra analysis into visual reports π
Hex is using GPT-6 Astra to turn complex analysis into visual reports. The useful bit is not just asking a model to summarize a table. It is moving from an analysis request to a result people can inspect and share.
What this points to:
β’ Analysis as an artifact: teams need charts, assumptions, and outputs, not a chat answer pasted into Slack.
β’ Human review still matters: a clean report can hide bad joins, stale data, or a wrong metric definition.
β’ Tool context is the product: models get more useful when they operate inside the workspace where data and business logic already live.
For AI app builders, this is the bar: produce a result that can survive review, not just a plausible paragraph.
[ Read More ] :
https://openai.com/index/hex-gpt-6-astra
γ°γ°γ°γ°γ°γ°
#AI #LLM #GPT6 #Data
@ProgrammingTip
Hex is using GPT-6 Astra to turn complex analysis into visual reports. The useful bit is not just asking a model to summarize a table. It is moving from an analysis request to a result people can inspect and share.
What this points to:
β’ Analysis as an artifact: teams need charts, assumptions, and outputs, not a chat answer pasted into Slack.
β’ Human review still matters: a clean report can hide bad joins, stale data, or a wrong metric definition.
β’ Tool context is the product: models get more useful when they operate inside the workspace where data and business logic already live.
For AI app builders, this is the bar: produce a result that can survive review, not just a plausible paragraph.
[ Read More ] :
https://openai.com/index/hex-gpt-6-astra
γ°γ°γ°γ°γ°γ°
#AI #LLM #GPT6 #Data
@ProgrammingTip
OpenAI
Hex turns complex analysis into visual reports with GPTβ6 Astra
GPT-6 Astra helps Hexβs data agents turn answers into interactive visualizations that employees are proud to share.
GPT-6 Sol and GPT-6 Luna are in the API π
OpenAI shipped GPT-6 Sol and GPT-6 Luna for API developers, alongside their availability in Codex and ChatGPT.
This is a two-model release, so do not blindly swap your existing production model. Put both behind the same eval set first: tool calls, structured output, long-context retrieval, refusal behavior, and latency under your real prompt size.
What to do this week:
β’ Add Sol and Luna as versioned model options in your config.
β’ Run replay traffic against a fixed golden set.
β’ Log model ID, token use, tool errors, and task success separately.
A model migration is an engineering change, not a dropdown change.
[ Read More ] :
https://openai.com/index/introducing-gpt-6-sol-and-luna/
γ°γ°γ°γ°γ°γ°
#AI #OpenAI #LLM #API
@ProgrammingTip
OpenAI shipped GPT-6 Sol and GPT-6 Luna for API developers, alongside their availability in Codex and ChatGPT.
This is a two-model release, so do not blindly swap your existing production model. Put both behind the same eval set first: tool calls, structured output, long-context retrieval, refusal behavior, and latency under your real prompt size.
What to do this week:
β’ Add Sol and Luna as versioned model options in your config.
β’ Run replay traffic against a fixed golden set.
β’ Log model ID, token use, tool errors, and task success separately.
A model migration is an engineering change, not a dropdown change.
[ Read More ] :
https://openai.com/index/introducing-gpt-6-sol-and-luna/
γ°γ°γ°γ°γ°γ°
#AI #OpenAI #LLM #API
@ProgrammingTip
OpenAI
Introducing GPT-6 Sol and Luna
Meet GPT-6 Sol and Luna, two models that bring frontier intelligence to everyday work with different balances of capability and cost.
Claude Opus 5.5 is now on the Claude Platform π
Anthropic introduced Claude Opus 5.5, a lower-cost frontier model available in Claude Code and through the Claude Platform.
That matters if your agent workload has been split between a high-end model for hard tasks and cheaper models for everything else. A lower-cost Opus tier can change where that handoff happens.
What to check:
β’ Run your existing eval set, not a few cherry-picked prompts.
β’ Measure tool-call accuracy and recovery after a failed call.
β’ Compare total agent cost: tokens, retries, and human review time.
For Claude Code users, model choice is now a practical repo-level config decision, not just a benchmark chart.
[ Blog ] :
https://www.anthropic.com/claude-opus-5-5
γ°γ°γ°γ°γ°γ°
#AI #LLM #Claude #ClaudeCode
@ProgrammingTip
Anthropic introduced Claude Opus 5.5, a lower-cost frontier model available in Claude Code and through the Claude Platform.
That matters if your agent workload has been split between a high-end model for hard tasks and cheaper models for everything else. A lower-cost Opus tier can change where that handoff happens.
What to check:
β’ Run your existing eval set, not a few cherry-picked prompts.
β’ Measure tool-call accuracy and recovery after a failed call.
β’ Compare total agent cost: tokens, retries, and human review time.
For Claude Code users, model choice is now a practical repo-level config decision, not just a benchmark chart.
[ Blog ] :
https://www.anthropic.com/claude-opus-5-5
γ°γ°γ°γ°γ°γ°
#AI #LLM #Claude #ClaudeCode
@ProgrammingTip
Anthropic
Introducing Claude Opus 5.5
Claude Opus 5.5 leads in agentic coding and knowledge work, and costs 40% less to run than Opus 5 on typical workloads.
Claude Opus 5.5 is available in the API π
Anthropic released Claude Opus 5.5, and it is available through the Claude API.
The practical pitch is simple: lower typical token costs than Opus 5, plus a faster mode when response time matters more than squeezing out the last bit of reasoning.
What to check in your evals:
β’ Run the same tool-use and coding tasks against Opus 5.
β’ Measure latency separately for normal and faster mode.
β’ Track input and output tokens, not just the model's listed price.
A cheaper high-end model changes agent architecture decisions. Some workflows that needed routing to a smaller model may now fit under one stronger default.
[ Read More ] :
https://www.anthropic.com/claude/opus
γ°γ°γ°γ°γ°γ°
#AI #LLM #Claude #API
@ProgrammingTip
Anthropic released Claude Opus 5.5, and it is available through the Claude API.
The practical pitch is simple: lower typical token costs than Opus 5, plus a faster mode when response time matters more than squeezing out the last bit of reasoning.
What to check in your evals:
β’ Run the same tool-use and coding tasks against Opus 5.
β’ Measure latency separately for normal and faster mode.
β’ Track input and output tokens, not just the model's listed price.
A cheaper high-end model changes agent architecture decisions. Some workflows that needed routing to a smaller model may now fit under one stronger default.
[ Read More ] :
https://www.anthropic.com/claude/opus
γ°γ°γ°γ°γ°γ°
#AI #LLM #Claude #API
@ProgrammingTip
Anthropic
Claude Opus
Hybrid reasoning model built for serious coding and AI agents, featuring a 1M context window.
Claude Sonnet 5.5 is built for coding agents π
Anthropic released Claude Sonnet 5.5 for the Claude Platform. It targets the work developers actually hand to coding agents: navigating a repo, making changes across files, using tools, and checking the result.
What changed:
β’ Faster agentic work: aimed at shorter tool loops and less idle time while an agent investigates a codebase.
β’ Lower-cost option: positioned for teams that need to run coding tasks repeatedly, not just ask one-off questions.
β’ Production focus: Anthropic calls out software engineering and multi-step agent workflows directly.
If your agent spends more time calling tools than writing code, model latency and per-task cost matter as much as benchmark scores. Test it on a real issue queue, with your actual tool permissions.
[ Read More ] :
https://www.anthropic.com/claude-sonnet-5-5
γ°γ°γ°γ°γ°γ°
#AI #LLM #Claude #CodingAgents
@ProgrammingTip
Anthropic released Claude Sonnet 5.5 for the Claude Platform. It targets the work developers actually hand to coding agents: navigating a repo, making changes across files, using tools, and checking the result.
What changed:
β’ Faster agentic work: aimed at shorter tool loops and less idle time while an agent investigates a codebase.
β’ Lower-cost option: positioned for teams that need to run coding tasks repeatedly, not just ask one-off questions.
β’ Production focus: Anthropic calls out software engineering and multi-step agent workflows directly.
If your agent spends more time calling tools than writing code, model latency and per-task cost matter as much as benchmark scores. Test it on a real issue queue, with your actual tool permissions.
[ Read More ] :
https://www.anthropic.com/claude-sonnet-5-5
γ°γ°γ°γ°γ°γ°
#AI #LLM #Claude #CodingAgents
@ProgrammingTip
Anthropic
Introducing Claude Sonnet 5.5
Claude Sonnet 5.5 is a clear upgrade over Claude Sonnet 5, runs 30%+ faster, and costs up to 30% less for most work.
Cloudflare shipped an agentic CLI for its API β‘οΈ
Cloudflare launched
This is a practical alternative to collecting one-off curl commands in a wiki or maintaining small admin scripts for every service. The CLI gives humans and coding agents one command-line entry point for Cloudflare operations.
Where it fitsβ
:
β’ Inspect and change Cloudflare resources while debugging a service.
β’ Give an agent a constrained operational interface instead of raw dashboard access.
β’ Turn repeatable incident steps into checked-in commands and runbooks.
Do not hand an agent broad production credentials because it has a nice CLI. Use scoped tokens, separate environments, and audit the resulting changes.
[ Blog ] :
https://blog.cloudflare.com/cloudflare-cf-cli-launch
γ°οΈγ°οΈγ°οΈγ°οΈγ°οΈγ°οΈ
#Cloudflare #CLI #DevOps #Agents #LLM
@ProgrammingTip
Cloudflare launched
cf, an agentic CLI for its full API surface.This is a practical alternative to collecting one-off curl commands in a wiki or maintaining small admin scripts for every service. The CLI gives humans and coding agents one command-line entry point for Cloudflare operations.
Where it fits
β’ Inspect and change Cloudflare resources while debugging a service.
β’ Give an agent a constrained operational interface instead of raw dashboard access.
β’ Turn repeatable incident steps into checked-in commands and runbooks.
Do not hand an agent broad production credentials because it has a nice CLI. Use scoped tokens, separate environments, and audit the resulting changes.
[ Blog ] :
https://blog.cloudflare.com/cloudflare-cf-cli-launch
γ°οΈγ°οΈγ°οΈγ°οΈγ°οΈγ°οΈ
#Cloudflare #CLI #DevOps #Agents #LLM
@ProgrammingTip
Please open Telegram to view this post
VIEW IN TELEGRAM
Cloudflare Blog
Introducing cf: the agentic CLI for the entire Cloudflare API
We are releasing cf, our new command-line tool that mirrors the entire Cloudflare API and supports programmatic TypeScript configuration. We are also open-sourcing Forge, our internal SDK generator.
GPT-6 Astra gets an Ultrafast API tier π
OpenAI added an Ultrafast speed tier for GPT-6 Astra in Codex and the API.
The important bit is real-time work. Astra can now run Responses API workflows over WebSockets, which is a much better fit for interactive coding assistants, live agent status, and UI flows where waiting on a full request feels bad.
What to check:
β’ Responses API: use it for the agent loop and tool calls.
β’ WebSockets: keep one live connection instead of polling.
β’ Ultrafast tier: test it where latency matters more than squeezing every last token of quality.
If your app streams agent work to a browser, this is worth benchmarking against your current model setup.
[ Article ] :
https://community.openai.com/t/build-ultrafast-with-astra-in-codex-and-the-api/1402393
γ°γ°γ°γ°γ°γ°
#AI #OpenAI #API #LLM
@ProgrammingTip
OpenAI added an Ultrafast speed tier for GPT-6 Astra in Codex and the API.
The important bit is real-time work. Astra can now run Responses API workflows over WebSockets, which is a much better fit for interactive coding assistants, live agent status, and UI flows where waiting on a full request feels bad.
What to check:
β’ Responses API: use it for the agent loop and tool calls.
β’ WebSockets: keep one live connection instead of polling.
β’ Ultrafast tier: test it where latency matters more than squeezing every last token of quality.
If your app streams agent work to a browser, this is worth benchmarking against your current model setup.
[ Article ] :
https://community.openai.com/t/build-ultrafast-with-astra-in-codex-and-the-api/1402393
γ°γ°γ°γ°γ°γ°
#AI #OpenAI #API #LLM
@ProgrammingTip
OpenAI Developer Community
Build Ultrafast with Astra in Codex and the API
Our premium speed tier, Ultrafast offers up to 8x faster token generation (300 tokens per second) in Codex and up to 6x in the API. In Codex, Ultrafast generates code faster, so you can move more quickly from idea to code to iteration π β¦
OpenAI Decisions API is in public beta π
OpenAI's Decisions API is now in public beta. It uses GPT-6 Luna to turn text or image input into structured decision results instead of another blob of prose.
What it returns:
β’ Predicates: true or false checks for policy, routing, or eligibility.
β’ Choices: select one option from a defined set.
β’ Scores: rank or grade an input on a scale you control.
This is useful when an LLM sits inside a real workflow. Send the result to an approval queue, a feature flag, or a C# switch expression. Keep the model's job narrow, validate the typed output, and do not make your app parse "I think yes".
[ Read More ] :
https://developers.openai.com/api/docs/guides/decisions
γ°γ°γ°γ°γ°γ°
#AI #OpenAI #LLM #API
@ProgrammingTip
OpenAI's Decisions API is now in public beta. It uses GPT-6 Luna to turn text or image input into structured decision results instead of another blob of prose.
What it returns:
β’ Predicates: true or false checks for policy, routing, or eligibility.
β’ Choices: select one option from a defined set.
β’ Scores: rank or grade an input on a scale you control.
This is useful when an LLM sits inside a real workflow. Send the result to an approval queue, a feature flag, or a C# switch expression. Keep the model's job narrow, validate the typed output, and do not make your app parse "I think yes".
[ Read More ] :
https://developers.openai.com/api/docs/guides/decisions
γ°γ°γ°γ°γ°γ°
#AI #OpenAI #LLM #API
@ProgrammingTip
OpenAI Developers
Decisions | OpenAI API
Use the Decisions API to check conditions, select from fixed options, and score text, images, and JSON against a rubric.
Claude Haiku 5.5 is available on the Claude Platform π
Anthropic released Claude Haiku 5.5, its lower-cost small model, on the Claude Platform.
The model ID is
What to test first:
β’ Replace your fast-path model in staging.
β’ Measure tool-call accuracy, not just chat quality.
β’ Set a harder-model fallback for failed evaluations.
Small models pay off when they handle the boring 80 percent cleanly. Keep the expensive model for the weird tickets.
[ Read More ] :
https://www.anthropic.com/claude-haiku-5-5
γ°γ°γ°γ°γ°γ°
#AI #LLM #Claude #API
@ProgrammingTip
Anthropic released Claude Haiku 5.5, its lower-cost small model, on the Claude Platform.
The model ID is
claude-haiku-5-5. This is the one to benchmark when your workload has lots of short agent turns: classification, extraction, routing, simple tool calls, and coding tasks where latency matters more than maximum reasoning depth.What to test first:
β’ Replace your fast-path model in staging.
β’ Measure tool-call accuracy, not just chat quality.
β’ Set a harder-model fallback for failed evaluations.
Small models pay off when they handle the boring 80 percent cleanly. Keep the expensive model for the weird tickets.
[ Read More ] :
https://www.anthropic.com/claude-haiku-5-5
γ°γ°γ°γ°γ°γ°
#AI #LLM #Claude #API
@ProgrammingTip
Anthropic
Introducing Claude Haiku 5.5
Claude Haiku 5.5 is our fastest, most capable small model. Built for high-volume work like summarization, subagents, and browser use.
Give LLM tasks a restart point π‘
Long agent prompts fail badly when the only recovery path is "start over."
Split work into stages that produce an artifact your next call can inspect: a JSON plan, a file list, a patch, test output, or a short decision log. Store that artifact outside the chat context.
Good loop:
β’ Ask for a plan with explicit acceptance checks.
β’ Run tools or tests yourself.
β’ Send back the actual output, not "it failed."
β’ Ask for the smallest next change.
This also makes retries cheaper. If call three fails, rerun call three with the saved result from call two. Do not pay for the model to rediscover the repository and your constraints every time.
γ°γ°γ°γ°γ°γ°
#AI #LLM #Agents #DevTools
@ProgrammingTip
Long agent prompts fail badly when the only recovery path is "start over."
Split work into stages that produce an artifact your next call can inspect: a JSON plan, a file list, a patch, test output, or a short decision log. Store that artifact outside the chat context.
Good loop:
β’ Ask for a plan with explicit acceptance checks.
β’ Run tools or tests yourself.
β’ Send back the actual output, not "it failed."
β’ Ask for the smallest next change.
This also makes retries cheaper. If call three fails, rerun call three with the saved result from call two. Do not pay for the model to rediscover the repository and your constraints every time.
γ°γ°γ°γ°γ°γ°
#AI #LLM #Agents #DevTools
@ProgrammingTip
Claude Fable 5.2 and Mythos 5.2 are out π
Anthropic shipped Claude Fable 5.2 and Mythos 5.2.
The useful part for teams is simple: model swaps are not just a leaderboard event. They change what your eval suite needs to catch before you update a production alias.
Before switching:
β’ Run your real tool-call and structured-output cases.
β’ Check retry rates, latency, and token use.
β’ Keep a small regression set for the weird customer inputs.
A new model can write cleaner code and still break one JSON edge case that your old one handled. Pin versions where you can, test the candidate, then move traffic deliberately.
[ Article ] :
https://www.anthropic.com/claude-fable-and-mythos-5-2
γ°γ°γ°γ°γ°γ°
#AI #LLM #Claude
@ProgrammingTip
Anthropic shipped Claude Fable 5.2 and Mythos 5.2.
The useful part for teams is simple: model swaps are not just a leaderboard event. They change what your eval suite needs to catch before you update a production alias.
Before switching:
β’ Run your real tool-call and structured-output cases.
β’ Check retry rates, latency, and token use.
β’ Keep a small regression set for the weird customer inputs.
A new model can write cleaner code and still break one JSON edge case that your old one handled. Pin versions where you can, test the candidate, then move traffic deliberately.
[ Article ] :
https://www.anthropic.com/claude-fable-and-mythos-5-2
γ°γ°γ°γ°γ°γ°
#AI #LLM #Claude
@ProgrammingTip
Anthropic
Introducing Claude Fable 5.2 and Claude Mythos 5.2
Claude Fable 5.2 and Claude Mythos 5.2, our new models for the hardest technical challenges. Fable 5.2 sets a new bar in coding, finance, and the sciences.