X keyword tripwire
Community · source page ↗ · last checked Oct 7, 2026, 6:03 PM
Current page contents
captured Oct 7, 2026, 6:03 PM2026-10-07T18:03:14.311Z
What we last captured from the live page ↗. ChangeRadar versions this and records what changed over time.
Change history
- Signal Oct 7, 2026, 2:01 PM
GPT-6.1 Sol launched at OpenAI DevDay: near GPT-6 Astra quality at $2/$10 per M tokens vs $10/$50, with 1.05M context. A 5x price cut at similar capability changes which agent workloads are worth running.
- Signal Oct 7, 2026, 12:59 PM
OpenAI shuts down gpt-4, gpt-3.5-turbo, o1 and o3-mini on Oct 23. Is your code still using them? One command: pipx run llm-sunset → every deprecated model ID in your repo, its shutdown date, and the replacement. Free, open source, works in CI. https://t.co/NHpuJLbJYs
- Signal Oct 7, 2026, 12:15 PM
New: Claude Opus 5.5 is cheaper and faster than Opus 5. For teams running agents in production, the pricing change matters more than the benchmark gains Anthropic cut Opus pricing 20-60% across the board and reported real cod https://t.co/jJ8np4HY2Q #AIagents #ARK
- Signal Oct 7, 2026, 11:50 AM
Grok 4.7 launched at $2 per million input tokens — half the price of GPT-5.6 Sol. Everyone's celebrating the price cut. Almost nobody's reading the second page. 🧵 https://t.co/YU6bzmBWKi
- Signal Oct 7, 2026, 10:46 AM
🍌 Google dropped Nano Banana 2.1 — 4K image gen/edit, handles 14 reference images, keeps 4 characters / 10 objects consistent. API price cut ~50% to $0.0336/1K images. Rolling out to Gemini app and AI Studio. (THE DECODER) https://t.co/KznseegnUm #Google #AI
- Signal Oct 6, 2026, 9:42 AM
GPT-6.1 Sol with cached input at 95% discount is the kind of pricing change that actually changes my architecture. I moved a whole pipeline to cached prompts last month and the bill dropped by half. Cached input is the real feature. https://t.co/OdxjyLELtu
- Signal Oct 5, 2026, 11:46 AM
"Introducing GPT-6.1 Sol Ultrafast. Our highest tier subscribers can now use Ultrafast for our newest Sol model to restore it's token speed back to it's unnerfed state at just a 6x price increase. Access will roll out across the next few days." https://t.co/HDEgJNpQqm
- Signal Oct 4, 2026, 9:09 PM
Check DeepSeek's new pricing page and you'll find a line built for a kind of customer that barely exists yet at most labs: off-peak hours, Monday through Friday, cached tokens running at $0.003 per million, half of peak. That's the headline nobody's writing, because everyone's https://t.co/i55UyofaaG
- Signal Oct 4, 2026, 5:18 PM
📢 https://t.co/iphkIDCIXD OCTOBER 4 DISCOUNT UPDATE A new pricing matrix is now in effect: 🔹 MiMo-V2.6-Flash → 90% off 🔹 Qwen3.8-Flash → 70% off 🔹 MiMo-V2.6-Pro → 50% off 🔹 DeepSeek-V4.1-Flash → 40% off 🔹 GLM-5.3-Flash → 30% off The biggest change is on DeepSeek and https://t.co/mWOz5eGw5u
- Signal Oct 4, 2026, 12:19 PM
40% less to run than Opus 5. Claude Opus 5.5 is here. and Anthropic says it performs at the level of Fable 5.1. Opus 5.5 is the first model in Anthropic’s new Claude 5.5 family. Anthropic says it generates output more than 30% faster than Opus 5. API pricing is $4M input and https://t.co/5PiCI2CkR1 https://t.co/caXrFkerQO
- Signal Oct 4, 2026, 11:28 AM
1/ OpenAI's GPT-6.1 Sol "nearly matches" GPT-6 Astra on agentic coding, computer use and professional work, according to OpenAI. At one-fifth of Astra's price. Not a new flagship. Not an Astra price cut. A cheaper mid-tier model that gets close 🧵 https://t.co/bNthk1XaJs
- Signal Oct 4, 2026, 8:59 AM
ChatGPT Pro still costs $200. From Oct 30, it gives 10x Plus usage, not 20x. The price stays the same. The usage drops by half. That is a price increase.
- Signal Oct 4, 2026, 7:39 AM
OpenAI just quietly changed their subscription pricing, and almost nobody noticed the fine print: • Pro 500 Plan: $500/month for 25x usage and "Ultrafast" mode (clocking 300 tokens/sec on Astra) • Ultrafast API: $60 per million input tokens, $300 per million output tokens •
- Signal Oct 4, 2026, 1:12 AM
Anthropic notified developers on Sep 30, 2026 that Claude Sonnet 4.5 will retire on the Claude API. Deprecated model: claude-sonnet-4-5-20250929 Tentative retirement: November 30, 2026 Recommended replacement: claude-sonnet-5-5
- Signal Oct 4, 2026, 12:01 AM
Anthropic has announced the deprecation of the Claude Sonnet 4.5 model, identified as claude-sonnet-4-5-20250929, with its retirement from the Claude API scheduled for November 30, 2026. This change directly impacts any production or development environments currently pinned to
- Signal Oct 3, 2026, 3:05 AM
Grok Now Answers Inside XChat Group Chats Group members can get replies by simply asking Grok in the chat. It is a feature inside X, not a new model. 📊 What changed: → Where: XChat group chats → How: ask Grok directly in the thread → Type: in-app feature, not a model launch
- Signal Oct 3, 2026, 2:06 AM
The DGX Spark has gone up in price, and that is the bad news. At a new price of $7000, alot of people are going to ask - can you do much with just 1 DGX Spark? The good news is you can run more frontier-class stuff on it than ever. 60-70 tok/s (322 in concurrency) on Qwen 3.8 https://t.co/khKoKDrym4 https://t.co/rkw0fX4e4p
- Signal Oct 2, 2026, 9:16 PM
DeepSeek made its 75% V4 Pro price cut permanent. Intelligence is becoming a commodity faster than anyone predicted. The winners won't be model builders — they'll be whoever routes compute like a portfolio. https://t.co/qAZfP7aw0V
- Signal Oct 2, 2026, 8:43 PM
GPT-6.1 Sol (Max) Hits #5 in Agent Arena at $0.56 Per Task GPT-6.1 Sol (Max) costs 88% less than Agent Arena's #1 model and trails it by only 3.08 pts. Arena says OpenAI's newcomer debuts at #5 with +11.23% net improvement vs baseline, at a $0.56 median cost per task. It now
- Signal Oct 2, 2026, 5:00 PM
Google just started sunsetting Gemini Gems in favor of Skills, migration begins in November. If you built workflows as Gems, this comparison is the one to read before they convert automatically. 💬 Comment GEMINI for a masterclass invite! https://t.co/qWzLIR0iGf
- Signal Oct 2, 2026, 12:30 PM
Claude Sonnet 5.5 keeps the same token prices as Sonnet 5. Anthropic says tasks cost up to 30% less because it uses fewer tokens. That is a claim about efficiency, not a price cut. Worth checking against your own workload. https://t.co/fGPVATO4xF
- Signal Oct 2, 2026, 2:35 AM
Claude Opus 5.5, GPT-6 Sol and Luna, and Grok 4.7 all shipped new pricing within 48 hours. Here's what each company's own pricing page actually says. https://t.co/RGf1kRg5io
- Signal Sep 30, 2026, 10:42 PM
gemini 4 argon's cost numbers on artificial analysis use its introductory pricing. AA lists argon at $1.99 per task. that's calculated with the current $2/$10 per million tokens. but that's temporary. once the introductory period ends, argon goes up tp $4/$20. the benchmark https://t.co/PHwwIp3073
- Signal Sep 30, 2026, 9:54 PM
Google is officially sunsetting Gemini Gems and replacing them with a more versatile tool called Skills. https://t.co/ptNZwLYJK5
- Signal Sep 30, 2026, 9:12 PM
Anthropic's Claude platform deprecation page added a November 30, 2026 retirement notice: "claude-sonnet-4-5- | claude-sonnet-5-5". The same edit changed "Retirement date (table 10)" from "null" to "Deprecated model | Recommended replacement". #software https://t.co/ZHVAlx5SP2
- Signal Sep 30, 2026, 5:47 PM
you don't need to overthink AI pricing strategy 💡 two companies. two approaches. both repriced 2026. Anthropic: mandatory enterprise metered billing OpenAI: 80% price cut on Luna (cheapest tier) here's what happened: →… tomtunguz.com/anthropic-repriced-the-enterprise/
- Signal Sep 30, 2026, 5:02 PM
OpenAI said in June it was sunsetting ChatGPT Pulse, about nine months after launch. For Pro users it did research overnight and had five to ten briefs ready by morning. They sat in a tab inside the ChatGPT app. Overnight work is only as useful as its address. The official line https://t.co/1A4HlmsCyJ
- Signal Sep 30, 2026, 3:14 PM
AI pricing can change quickly, but users do not necessarily have to lose access when they do. That distinction is what makes the latest DeepSeek-V4-Flash update on https://t.co/JGWo7qKovy interesting. The official price has increased, yet according to the announcement, https://t.co/PIdWnTxV4l
- Signal Sep 30, 2026, 3:06 PM
AI pricing can change quickly, but users do not necessarily have to lose access when they do. That distinction is what makes the latest DeepSeek-V4-Flash update on https://t.co/VEBWyGT89E interesting. The official price has increased, yet according to the announcement, https://t.co/ogNqDSumxG
- Signal Sep 30, 2026, 2:02 PM
Why Claude Sonnet 5.5 Is Cheaper Without a Price Cut, and When It Isn’t https://t.co/GuxoXdfpWE
- Signal Sep 30, 2026, 1:38 PM
devday was mostly a price cut. gpt-6.1 sol at one fifth of astra's token price, computer use in the agents api, codex taking voice commands now
- Signal Sep 28, 2026, 4:51 PM
Google is officially killing Gemini Gems -Google has officially announced that it’s sunsetting Gemini Gems. -It will start migrating Gems to Skills on November 17. -It’s unclear if users without a Google AI subscription will be able to access Skills.
- Signal Sep 28, 2026, 1:53 PM
The AI price war just went nuclear — and it's the best news for builders all year. On September 22, Anthropic shipped Claude Opus 5.5: $4 per million input tokens, $20 per million output — a 20% price cut, cache reads 60% cheaper, and the five-hour usage cap gone for Pro, Max, https://t.co/H9omUahyhr
- Signal Sep 28, 2026, 11:01 AM
The most useful AI release last week wasn't a frontier model. It was a price cut. Alibaba shipped Qwen-Audio 3.1 on 23 September. Speech recognition up to 95% cheaper. Realtime voice about 85% cheaper. Text to speech about 70% cheaper. If you build for Indian businesses, that
- Signal Sep 28, 2026, 1:28 AM
🎙️ DeepSeek-V4.1-Flash: https://t.co/7OU6BjdxLB Stacks an Extra 90% Off 🚀💰 DeepSeek has launched its new native multimodal MoE model, DeepSeek-V4.1-Flash. At the same time, https://t.co/7OU6BjdxLB is introducing an additional 90% discount on top of DeepSeek's new official https://t.co/l98JRuGJGV https://t.co/YFMF5xsMk0
- Signal Sep 28, 2026, 1:23 AM
🌗 https://t.co/7OU6BjdxLB Prepares a “Stacked Surprise” for DeepSeek V4 Flash Series 👀⚡ DeepSeek-V4-Flash and DeepSeek-V4-Flash-Vision-Exp are approaching a pricing change on September 10 at 12:00 SGT. But instead of simply passing upstream pricing changes through to https://t.co/tRDqdO4ItC https://t.co/OW37TUR7VJ
- Signal Sep 27, 2026, 10:23 AM
for builders using DeepSeek — a price cut just got even bigger on https://t.co/dayONJgKXL. DeepSeek has officially launched its native multimodal MoE model, DeepSeek-V4.1-Flash. To celebrate the launch, https://t.co/dayONJgKXL is stacking an extra 90% OFF on top of DeepSeek's https://t.co/8p0xGLXqaC
- Signal Sep 26, 2026, 5:50 AM
DeepSeek raised its API prices last month. The Flash model went from $0.28 per million output tokens to $1.20 at peak, $0.60 off-peak. Everyone expected developers to leave for something cheaper. Instead its run rate doubled to over $1B, according to The Information, and the CEO https://t.co/P2czHaWaBn
- Signal Sep 26, 2026, 4:14 AM
📅 https://t.co/75nSdsZ6MN’S DEEPSEEK TRANSITION: TURN PRICING CHANGES INTO A PLAN https://t.co/75nSdsZ6MN’s announcement set September 3, 2026, at 17:00 SGT as the transition from free access to discounted pricing for DeepSeek V4 Flash and DeepSeek V4 Flash Vision Exp. The https://t.co/YnaaKr2od7
- Signal Sep 26, 2026, 2:03 AM
CLAUDE FABLE 5.1 IS NOW LIVE ON https://t.co/fEDs0n3D2S API! 📢 https://t.co/fEDs0n3D2S has added Claude Fable 5.1 from Anthropic, bringing a new model built for developers working with complex reasoning and long-horizon Agent workflows. Built on the same foundation as Claude https://t.co/28seNeVXcY https://t.co/spVdivTcFE
- Signal Sep 26, 2026, 1:19 AM
Voice AI just got a lot cheaper. Alibaba cut Qwen Audio API prices by up to 95%. ASR ↓ up to 95% Realtime ↓ ~85% TTS ↓ ~70% And they didn't just cut prices. Qwen Audio 3.1 adds: → realtime voice → function calling → web search → voice cloning → audio understanding → https://t.co/KeiIA3NB1n
- Signal Sep 25, 2026, 11:49 PM
Grok 4.7 shipped with 40% more parameters at the same price. On Terminal-Bench, it still scores 38%. Claude Opus 5.5 scores 66.4% on the same benchmark, at roughly 40% below what Opus 5 used to cost. The gap holds across the board. SWE-bench Pro: Opus 5.5 at 89.9%. CursorBench https://t.co/k6mXYpAV9m https://t.co/7bZLuLtIii
- Signal Sep 25, 2026, 7:33 PM
Anthropic just published Claude Opus 5.5's real-world task costs today. Here's what you need to know. Developer account ClaudeDevs broke down what the new pricing actually means for a typical Claude Code task, and built a calculator so users can run their own numbers from the https://t.co/WFkaiMSITw https://t.co/dmh3sYFA0A
- Signal Sep 25, 2026, 7:29 PM
Anthropic is on a generational PR run. - Opus 5.5 drops and is better than Fable and gets a price cut - gave every user a banked reset that doesn’t push back reset date - increased 5h session limits - free usage credit for cloud sessions - Sonnet 5.5 and Haiku 5.5 in the https://t.co/gNKFlVXXwU
- Signal Sep 24, 2026, 11:23 AM
gpt-6 launched today with a 50% api price cut. anthropic's already answering with opus 5.5. the actual move for operators: pull up whatever AI workflow you shelved last year because the per call cost didn't pencil out. redo the math this week. the floor just moved under you.
- Signal Sep 24, 2026, 9:29 AM
Your next default model gets picked on price, and it's moving fast. - Opus 5.5 is winning people over on limits and price this week - OpenAI shipped GPT-6 and says the 50% API price cut is permanent - Dan Shipper has a new daily driver contender that runs up to 40% cheaper than https://t.co/kNUHANKfUS
- Signal Sep 24, 2026, 9:07 AM
https://t.co/fEDs0n3D2S MODEL RIGHTS UPDATE: DEEPSEEK V4 FLASH PRICING CHANGES https://t.co/fEDs0n3D2S has announced an upcoming pricing update for DeepSeek V4 Flash and DeepSeek V4 Flash Vision Exp. Starting September 3, 2026 at 17:00 SGT, both models will end their free trial https://t.co/qGtOctaQKa https://t.co/qRIjbmRgEN
- Signal Sep 24, 2026, 8:03 AM
💥 BIG REVEAL: DEEPSEEK PRICE CUT + https://t.co/fEDs0n3D2S STACKS AN EXTRA 90% OFF! ⚡🤖📉💸🌐 A seismic price-to-performance breakthrough has hit the developer ecosystem! With DeepSeek officially releasing its native multimodal MoE https://t.co/eO1DFnRZ6J https://t.co/CNArh5wyab
- Signal Sep 24, 2026, 7:20 AM
Anthropic just dropped Claude Opus 5.5. Same flagship tier, price cut to $4/$20 per million tokens (was $5/$25), plus a 1M token context window. AI labs are now competing on price as hard as capability. Good news if you're building on these models.
- Signal Sep 24, 2026, 5:55 AM
anthropic and openai dropped new models on the same day this week. Opus 5.5 costs 40% less than Opus 5. GPT-6 Sol and Luna landed with a 50% API price cut. two labs cutting prices the same week isn't competition - it's a race to become the default. https://t.co/WCjJohyPlX
- Signal Sep 24, 2026, 5:29 AM
Anthropic just launched Claude Opus 5.5—and the focus is pure production economics. Instead of another marginal benchmark bump, this release brings top-tier Fable intelligence at a 40% price cut. Here are the specs you need to know: 🧵👇 https://t.co/USaBhNrWgb
- Signal Sep 23, 2026, 10:26 PM
Google priced Gemini 3.8 Flash TTS at $32.98 per million characters. That is up from the model it replaces, not a price cut. It tops Google's pronunciation robustness benchmark at 89.5%. TTS is billed per character. Your users hear minutes. #TTSPricing
- Signal Sep 23, 2026, 9:45 PM
opus 5.5 is out at $4 in / $20 out per million tokens, 20% under opus 5. anthropic isn't pitching it as smarter chat, they're pitching it as agentic coding and computer use. the price cut on the flagship tier is the actual news.
- Signal Sep 22, 2026, 8:00 AM
Understanding Claude Code's New Pricing for Third-Party Tools Anthropic introduced separate billing for Claude Code users accessing third-party development tools like OpenClaw, shifting from subscription-included usage to pay-as-you-go... Written for those just getting started.
- Signal Sep 22, 2026, 2:38 AM
🔥 BIG REVEAL: DEEPSEEK OFFICIAL PRICE ADJUSTMENT + https://t.co/fEDs0n3D2S STACKS AN EXTRA 90% OFF (10% OF OFFICIAL PRICE)! ⚡🚀🌐 A seismic price cut just landed across frontier compute! DeepSeek has officially released its native multimodal MoE foundation model, https://t.co/VOx0hZXTFe https://t.co/XOghKnczdg
- Signal Sep 22, 2026, 2:35 AM
🔥 BIG REVEAL: DEEPSEEK PRICE CUT + https://t.co/fEDs0n3D2S STACKS AN EXTRA 90% OFF! ⚡🚀🌐 A massive compute price drop just leveled up! DeepSeek has officially launched its native multimodal MoE engine, DeepSeek-V4.1-Flash. To celebrate this frontier milestone, https://t.co/aQHrRu9QDD https://t.co/CNArh5wyab
- Signal Sep 21, 2026, 7:31 PM
Grok 4.7 is officially released! 🚀🤖 • 2.1 Trillion MoE parameter architecture • 500,000-token native context window • Outright #1 on EEBench hardware design (64.0%) • 71.0% DeepSWE autonomous coding • 88% token price cut ($2/$6 per 1M) Full technical teardown in reply 👇 https://t.co/RUmE6ty9Wh
- Signal Sep 20, 2026, 10:55 PM
Alibaba’s Qwen3.8-Omni-Flash matches Gemini 3.8 Flash at 80%+ lower pricing ($0.15 in, $0.47 out per 1M tokens). Multimodal streaming went from a luxury spend to pennies. Does an 80% price cut actually change what you ship, or was cost never the real blocker?
- Signal Sep 19, 2026, 5:57 PM
Alibaba dropped Qwen3.8-Omni-Flash. Text/image/audio/video in. 1M context, and the best part: audio input price cut 98%. Three days after Gemini 3.8 Live owned cheap speech. Talk about a price/perf race, right? https://t.co/rpqS3Qyemw
- Signal Sep 19, 2026, 9:40 AM
💥 BIG REVEAL: DEEPSEEK PRICE CUT + https://t.co/fEDs0n35dk STACKS AN EXTRA 90% OFF! ⚡🚀 The ultimate foundation model price breakthrough has arrived! To celebrate DeepSeek officially launching its native multimodal MoE architecture, DeepSeek-V4.1-Flash, https://t.co/fEDs0n35dk https://t.co/odGVcA0Unl https://t.co/CNArh5w0kD