AICoder

AI Coding Live News

First-hand AI coding updates from 𝕏/Twitter frontier labs and researchers, distilled by AI with entity benchmarks.

Explore live on 𝕏 ↗

AI Coding Live News

O
OpenAI DevelopersOpenAIDevs·
👀61
🔥81
View on 𝕏 ↗

OpenAI Agents API ships Auto Review, async compaction, experimental time awareness and mid-session tool updates

The OpenAI Developers account shared a post by Steve (@stevendcoffey), who works on the API at OpenAI, listing five Agents API updates this week: Codex-style Auto Review of commands before execution, compaction that runs in the background, an experimental time-awareness feature, updating a session's tools at any time, and automatic caching of system prompts across all sessions. As of writing, the official API changelog does not list these items yet.

Key Takeaways

  • Five updates: Auto Review, async compaction, time awareness (experimental), updating session tools at any time, and automatic cross-session system prompt caching
  • Auto Review mirrors the Codex mechanism and evaluates commands for potential bad behavior before they run
  • Time awareness lets the model act on time-boxed instructions such as "timebox CI fixes to 10 minutes"
  • The Agents API entered public beta on September 10, and this is another round of features within 7 weeks (after computer use on Sep 29, sandbox prewarming on Oct 9 and Ultrafast mode on Oct 16)
  • OpenAI has published no latency or cost-saving numbers, and the changelog has not listed these items yet
Read full story
Linked Models & Agents:
🤖 Agents API🤖 Codex🌐 OpenAI
#OpenAI#Agents API#Codex#Auto Review#Compaction#prompt caching#Agents
O
OpenAI DevelopersOpenAIDevs·
👀83
🔥84
View on 𝕏 ↗

OpenAI API adds 12-hour prompt caching: GPT-5.6+ models can keep shared context cached for half a day, with writes at 2.5x input and reads still at 0.1x

On Oct 20 OpenAI added 12-hour prompt caching to the Responses API: GPT-5.6 and later models can set prompt_cache_options.ttl to "12h" so instructions, tool definitions and reference material stay billed at the cached rate even when requests are hours apart. 12-hour writes cost 2.5x the ordinary input price versus 1.25x for the default 30 minutes; read rates are unchanged at 0.1x on most models and 0.05x on GPT-6.1 Sol. One request can mix 12-hour and 30-minute breakpoints.

Key Takeaways

  • New option: prompt_cache_options.ttl in the Responses API accepts "30m" (default) or "12h", and prompt_cache_breakpoint.ttl overrides it per breakpoint, for GPT-5.6 and later models
  • Rates (vendor-stated): 30-minute writes at 1.25x and 12-hour writes at 2.5x ordinary input; reads at 0.1x on most models and 0.05x on GPT-6.1 Sol, same for both lifetimes
  • Our estimate from the documented rates: reusing one prefix 6 times an hour apart costs about 6x1.25=7.5x input with 30-minute caching (rewritten each time) versus about 2.5+5x0.1=3.0x with 12-hour caching
  • Up to 4 cache writes per request and a 1,024 visible-token minimum prefix; usage now splits writes into ttl_30m_tokens and ttl_12h_tokens
  • "12h" is a minimum lifetime, not a deletion deadline, and does not guarantee a hit or lower latency; GPT-5.5 and earlier keep using prompt_cache_retention
Read full story
Linked Models & Agents:
⚡ GPT-6.1 Sol⚡ GPT-5.6🌐 OpenAI
#Prompt Caching#Responses API#Agent#OpenAI
I
Inception_inception_ai·
👀238
🔥92
View on 𝕏 ↗

Inception launches Mercury 3 diffusion LLM: 1,068 tokens/s and 1M context, beating GPT-6 Luna and Haiku 5.5 on several benchmarks per the vendor

On October 19 Inception released Mercury 3, which it calls the most capable diffusion LLM (dLLM) to date. Per the vendor, it is more than 40x the size of Mercury 1, outputs 1,068 tokens/s, expands context from 260K (Mercury 2.5) to 1M tokens, and beats GPT-6 Luna and Claude Haiku 5.5 (both medium) on GPQA Diamond, IFBench, SWE-Bench Atlas and more. List price is $0.20/M input and $0.70/M output, 50% off at launch. It is live on the Inception API, OpenRouter, Baseten and Vercel AI Gateway with an OpenAI-compatible API. All scores are vendor-reported and not yet independently reproduced.

Inception launches Mercury 3 diffusion LLM: 1,068 tokens/s and 1M context, beating GPT-6 Luna and Haiku 5.5 on several benchmarks per the vendor

Key Takeaways

  • Speed: 1,068 tokens/s output; the vendor table lists GPT-6 Luna at ~130 and Haiku 5.5 at ~240 (vendor figures).
  • Scores: GPQA Diamond 89.23, IFBench 76.84, SWE-Bench Atlas 51.08, all above Luna and Haiku 5.5; but DeepSWE v1 at 46.31 trails Haiku 5.5 (51.62), and MRCR v2 also trails Haiku.
  • Scale and context: more than 40x the parameters of Mercury 1, with a 1M-token window (260K on Mercury 2.5).
  • Price: $0.20/M in and $0.70/M out list, 50% off at launch ($0.10/$0.35), cached input 90% off; new API keys get 100M free tokens.
  • Access: Inception API, OpenRouter, Baseten and Vercel AI Gateway, OpenAI-compatible as a drop-in replacement.
Read full story
Linked Models & Agents:
⚡ Mercury 3⚡ Mercury 2.5⚡ GPT-6 Luna⚡ Claude Haiku 5.5🌐 Inception🌐 OpenRouter🌐 Baseten🌐 Vercel
#Diffusion LLM#Mercury#Inception
O
OpenAI@OpenAI·
👀131
🔥87

GPT-5.5 retires from ChatGPT, ChatGPT Work and Codex today: ChatGPT sign-in users must move to gpt-6-sol or gpt-6-luna, the API is unaffected

Per OpenAI's model docs, GPT-5.5 retires from ChatGPT, ChatGPT Work and Codex on all plans on 2026-10-14; the OpenAI API is not affected. Codex users on ChatGPT sign-in should pick gpt-6-sol (Plus, Pro, Business, Enterprise, Edu) or gpt-6-luna in the desktop app (Free, Go), with gpt-6.1-sol recommended for complex coding. Any config.toml, custom agent, scheduled task or script pinned to gpt-5.5 needs updating now.

Key Takeaways

  • Effective 2026-10-14: GPT-5.5 leaves ChatGPT, ChatGPT Work and Codex on every plan; the OpenAI API is unaffected
  • Replacements: gpt-6-sol for Plus/Pro/Business/Enterprise/Edu, gpt-6-luna in the desktop app for Free/Go (both 'when available')
  • For complex coding OpenAI recommends gpt-6.1-sol, near-Astra at 1/5 of Astra's standard price per OpenAI
  • Check 4 places: model in config.toml, codex exec -m, custom agents, scheduled tasks and scripts
  • Enterprise/Edu admins must enable newer models first; GPT-5.6 Sol/Terra/Luna remain available during the rollout
Read full story
Linked Models & Agents:
⚡ GPT-5.5⚡ GPT-6 Sol⚡ GPT-6 Luna⚡ GPT-6.1 Sol🤖 Codex🌐 OpenAI
#OpenAI#Codex#ChatGPT#GPT-5.5#Deprecation#Migration#Gpt 6 sol#Gpt 6 luna#gpt-6.1-sol
ADSponsored
O
OpenAI@OpenAI·
👀131
🔥80

OpenAI API shuts down 12 legacy models on Oct 23: o4-mini, o3-mini, gpt-4o-2024-05-13 and gpt-image-1 included, 12 days left to migrate

Per OpenAI's official Deprecations page, a batch of legacy snapshots deprecated on Apr 22 stops serving in the API on Oct 23, 2026: 12 snapshots and their aliases including gpt-3.5-turbo-0125, gpt-4-0613, gpt-4-turbo, gpt-4.1-nano, gpt-4o-2024-05-13, gpt-image-1, o1, o1-pro, o3-mini and o4-mini, plus five fine-tuned model families. Official substitutes are gpt-5.6-sol/terra/luna and gpt-image-2.5. The #keep4o campaign to preserve GPT-4o snapshots resurfaced on X today.

Key Takeaways

  • 12 snapshots plus aliases go dark together: o4-mini, o3-mini, o1, o1-pro, gpt-4-turbo, gpt-4, gpt-3.5-turbo aliases stop resolving (OpenAI Deprecations)
  • Shutdown date Oct 23, 2026 — 12 days away; announced Apr 22, meeting OpenAI's 6-month minimum for GA models
  • Substitutes: o4-mini / gpt-3.5 → gpt-5.6-terra; gpt-4 family / o1 / o3-mini / gpt-4o-2024-05-13 → gpt-5.6-sol; o1-pro → gpt-5.6-sol with reasoning.mode: pro; gpt-4.1-nano → gpt-5.6-luna
  • gpt-image-1 → gpt-image-2.5-sunburst or -flare; five fine-tuned families (ft-gpt-4, ft-gpt-3.5-turbo, ft-o4-mini, etc.) also shut down and must be retrained on new bases
  • Next wave: Dec 1 (gpt-image-1.5 / 1-mini) and Dec 11 (first-gen gpt-5 snapshots, o3, o3-pro)
Read full story
Linked Models & Agents:
⚡ o4-mini⚡ OpenAI o3-mini (High Reasoning Coding)⚡ o1⚡ GPT 4o 2024.05.13⚡ GPT Image 1⚡ gpt-4.1-nano⚡ GPT-5.6 Sol⚡ GPT-5.6 Terra⚡ GPT-5.6 Luna⚡ gpt-image-2.5🌐 OpenAI
#OpenAI#API#O4 mini#GPT-4o#Gpt image 1
O
OpenAI@OpenAI·
👀156
🔥87

OpenAI opens the Decisions API public beta: POST /v1/decisions on gpt-6-luna returns typed predicate, choice and score answers ~10x faster than the Responses API, billed only for input at $0.10 per 1M tokens

Per the OpenAI API changelog, the Decisions API entered public beta on 2026-10-06 with gpt-6-luna: a dedicated POST /v1/decisions endpoint takes text/images plus a list of questions and returns typed predicate, choice or score answers with probabilities, about 10x faster than the Responses API. Billing is input-only at $0.10 per 1M tokens (no output or cache charges); ZDR and HIPAA are supported and GA is expected within weeks.

OpenAI opens the Decisions API public beta: POST /v1/decisions on gpt-6-luna returns typed predicate, choice and score answers ~10x faster than the Responses API, billed only for input at $0.10 per 1M tokens

Key Takeaways

  • Speed: typed answers about 10x faster than the Responses API (Decisions guide)
  • Price: input-only $0.10 per 1M tokens, no output or cache read/write fees; regular gpt-6-luna is $0.10 in / $0.50 out
  • 3 question types: predicate (0–1 probability), choice (value + probabilities + confidence), score (probability-weighted ordinal level)
  • Compliance: ZDR and HIPAA eligible, US and Europe (EEA + Switzerland) data residency; GA expected in weeks
  • Limits: gpt-6-luna only; images must be inline base64, no hosted URLs or file_id
Read full story
Linked Models & Agents:
⚡ GPT-6 Luna🌐 OpenAI
#OpenAI#Decisions API#Gpt 6 luna#Classification#Routing#Agents#API
C
ChatGPT@ChatGPT·
👀72
🔥92
View on 𝕏 ↗

OpenAI: GPT-5.5 leaves ChatGPT, Work, and Codex plans on Oct 14; switch to GPT-5.6 Sol or GPT-6 Astra

ChatGPT announced that on October 14, 2026, GPT-5.5 will leave ChatGPT, ChatGPT Work, and Codex across all plans; Codex users should switch to GPT-5.6 Sol or GPT-6 Astra. OpenAI Developers clarified that GPT-5.5 remains available via the OpenAI API Platform and in Codex sessions authenticated with an API key.

Key Takeaways

  • Oct 14: GPT-5.5 exits ChatGPT, ChatGPT Work, and Codex across all plans.
  • Codex plan users should move to GPT-5.6 Sol or GPT-6 Astra.
  • GPT-5.5 stays on the API Platform and in API-key-authenticated Codex sessions.
Read full story
Linked Models & Agents:
⚡ GPT-5.6 Sol⚡ GPT-6 Astra🤖 codex🌐 OpenAI
#GPT-5.5#GPT-5.6 Sol#GPT-6 Astra#ChatGPT#Codex#Deprecation#API
P
Pi Changelog@PiChangelog·
👀78
🔥88
View on 𝕏 ↗

Pi v0.85.1: GPT-6 Astra via OpenAI API keys and Codex; fixes 0.85.0 SDK import break

Pi Changelog ships v0.85.1: GPT-6 Astra is available via OpenAI API keys and OpenAI Codex subscriptions, and SDK import failures from 0.85.0 accidentally published experimental code are fixed. Experimental/plugin subpaths and server/client commands are now source-only via pi-test.sh; local SDK and stdio RPC API stay unchanged.

Key Takeaways

  • v0.85.1 unlocks GPT-6 Astra via OpenAI API keys and Codex subscriptions.
  • Fixes SDK import failures from 0.85.0 accidental experimental publish.
  • Experimental/plugin subpaths are source-only via pi-test.sh; local SDK and stdio RPC unchanged.
Read full story
Linked Models & Agents:
⚡ GPT-6 Astra🤖 pi🤖 codex🌐 OpenAI
#Pi#v0.85.1#GPT-6 Astra#OpenAI#Codex#SDK#Changelog
O
OpenAI@OpenAI·
👀48
🔥94
View on 𝕏 ↗

OpenAI: GPT-6 Astra rolling out to Plus/Pro/Enterprise, API, and AWS

OpenAI says GPT-6 Astra is rolling out today to a limited set of organizations and over coming days to all ChatGPT Plus, Pro, Business, and Enterprise users, plus the OpenAI API and AWS; it points users to the ChatGPT desktop app for the best experience.

Key Takeaways

  • Limited orgs today; Plus/Pro/Business/Enterprise over the coming days.
  • Also available via OpenAI API and AWS for developers.
  • Official tip: use the ChatGPT desktop app for best results.
Read full story
Linked Models & Agents:
🌐 OpenAI
#GPT-6 Astra#OpenAI#API#AWS#ChatGPT Plus#Enterprise#Rollout
O
OpenAI@OpenAI·
👀51
🔥89
View on 𝕏 ↗

OpenAI Previews Ultrafast Mode: GPT-5.6 Sol at up to 14x Speed

OpenAI previewed Ultrafast mode for GPT-5.6 Sol, powered by Cerebras, generating up to 750 tokens per second—as much as 14x faster. It launches first in the OpenAI API for a select group of customers, targeting latency-sensitive workflows such as real-time voice, commerce, coding, financial research, and security response.

Key Takeaways

  • GPT-5.6 Sol Ultrafast hits up to 750 tokens/sec, about 14x faster
  • Inference is powered by Cerebras for latency-critical products
  • Rolling out first via the OpenAI API for coding, voice, commerce, and security work
Read full story
Linked Models & Agents:
⚡ GPT-5.6 Sol🌐 OpenAI🌐 Cerebras
#AI#LLM#Coding#OpenAI#Inference
ADSponsored
O
OpenAI@OpenAI·
👀47
🔥93
View on 𝕏 ↗

Ultrafast GPT-5.6 Sol preview: up to 14x speed and 750 tokens/sec for coding and realtime work

OpenAI previewed Ultrafast mode: GPT-5.6 Sol at up to 14x speed, powered by Cerebras, generating up to 750 tokens per second. It launches first in the OpenAI API for a select customer group, then expands as capacity grows. Target workflows include realtime voice and support, commerce, coding and design, financial research, and security response—frontier intelligence where every second counts.

Key Takeaways

  • Frontier coding quality at near-instant completion speed is now a product mode, not just a benchmark demo.
  • 750 tokens/sec is Cerebras-backed; capacity, not the model card, is the rollout bottleneck.
  • Access is invite/select-customer preview, so production adoption still waits on fleet growth.
Read full story
Linked Models & Agents:
⚡ GPT-4o🤖 copilot🌐 OpenAI
#Ultrafast#gpt-5.6-sol#Cerebras#Latency#Coding#API
O
OpenAI Developers@OpenAIDevs·
👀82
🔥96
View on 𝕏 ↗

GPT-5.4 and GPT-5.4 mini leave ChatGPT-signed-in Codex today; switch to GPT-5.6 Terra and Luna

As of August 31, 2026, GPT-5.4 and GPT-5.4 mini are no longer available in Codex for users signed in with ChatGPT. OpenAI's recommended replacements are gpt-5.6-terra and gpt-5.6-luna. Both retiring models remain on the OpenAI API and in Codex sessions authenticated with an API key. Teams should update workspace defaults, custom agents, and scheduled tasks pinned to the old IDs.

Key Takeaways

  • ChatGPT-signed-in Codex drops GPT-5.4 and GPT-5.4 mini today in favor of GPT-5.6 Terra and Luna.
  • API-key Codex sessions and the OpenAI API keep access to the retiring models.
  • Update workspace defaults, custom agents, and scheduled tasks before they fail on pinned model IDs.
Read full story
Linked Models & Agents:
⚡ GPT-5.4⚡ GPT-5.4 mini⚡ GPT-5.6 Terra⚡ GPT-5.6 Luna🌐 OpenAI
#OpenAI#Codex#Gpt 5 4#Gpt 5 6#Deprecation
O
OpenAI@OpenAI·
👀47
🔥93
View on 𝕏 ↗

OpenAI previews Ultrafast mode: GPT-5.6 Sol at up to 14x the speed

OpenAI previewed Ultrafast mode, saying GPT-5.6 Sol can run at up to 14x its usual speed. Powered by Cerebras, Ultrafast generates up to 750 tokens per second for workflows where latency is the product: real-time voice and support, commerce, coding and design, financial research, and security response. It launches first in the OpenAI API to a select group of customers, with broader business access as capacity grows.

Key Takeaways

  • Flagship coding model Sol is now being sold on near-real-time throughput, not just quality.
  • 750 tokens/second makes interactive coding agents, voice, and support workflows viable.
  • Access is still a limited API preview; capacity is the gate to broader rollout.
Read full story
Linked Models & Agents:
⚡ GPT-5.6 Sol🌐 OpenAI
#Gpt 5 6 sol#Ultrafast#Cerebras#Latency#Codex#API
You've caught up with all stories