Anthropic — Compare Product Launches

Compare product launches across companies

Every product starts at Week 1 on its first observed publication date on or after its tracking date, falling back to that tracking date when no data exists. This comparison start is inferred from tracked content.

Products
Display

Mentions · Cumulative

Normalized data through Oct 5, 2026. Each week spans seven days from that product’s comparison start. Hollow points mark partial weeks; they contain recorded data only.

View week-by-week tables

Product summary

Totals cover the displayed window, including its partial week. Averages, peaks, and latest changes use completed weeks only. Products may have different amounts of completed history.

ProductComparison startCompleted weeksPartial progress MentionsItemsDistinct companiesAvg mentions / completed weekPeak completed weekLatest completed-week change
Opus 5.5 Sep 22, 2026
First observed data
2 — 978540 48.5 Week 1
83 mentions
-83.1%

Source breakdown

ProductX / Twitter
MentionsItemsDistinct companies
Opus 5.5978540

Week-by-week detail

Opus 5.5
Opus 5.5 · Mentions · Cumulative
Launch weekActual datesStatusMentionsItemsDistinct companiesCumulative mentionsMention growth
Week 1Sep 22, 2026 – Sep 28, 2026 Completed 83754083 —
Week 2Sep 29, 2026 – Oct 5, 2026 Completed 1410897 -83.1%

Recent content

Latest 50 matching items across selected products and their displayed windows. Items mentioning several products appear once.

ContentProductsSourcePublisherDate
RT @bentlegen: I wanted to know how far you can push these vids So I gave Opus 5.5 Blender, ElevenLabs and REAPER (DAW) and asked it to ma…Opus 5.5X / TwitterModemOct 5, 2026
RT @rseroter: Updated @antigravity in my personal Google AI Pro subscription. Well hello Claude Opus 5.5 and Sonnet 5.5. Pretty good value…Opus 5.5X / TwitterGoogle CloudOct 3, 2026
Google announced Gemini 4 Argon on September 30. - Vendor benchmarks: leads DeepSWE v1.1 and the Vals Index. - Artificial Analysis independent score: 52.6 vs 57.6 for Claude Opus 5.5. - Access: Fairwind cyber defenders only. No public API yet. Strong, uneven, and unavailable to most teams. Full breakdown → https://t.co/mODrd9gGL2 #Gemini4 #AIModels #AISecurity #CISOOpus 5.5X / TwitterNeuralTrustOct 2, 2026
This video was made with Claude Opus 5.5 and goes to show what TestMu AI does. Launched by Anthropic on September 22, 2026, Opus 5.5 is what the company calls "the strongest-performing model we've tested to date" built for agentic coding, knowledge work, computer use and multidisciplinary reasoning. You can use Claude Opus 5.5 to build and run video agents that plan, storyboard, edit, and render multimedia content autonomously. 🚀 Major perks of Opus 5.5: - Faster and cheaper to run than Opus 5 - StrongerOpus 5.5X / TwitterTestMu AIOct 2, 2026
The biggest generational jump since June came from a model that costs $0.18 per million output tokens. DeepSeek-V4-Flash-0731 (@Deepseek_AI) scored 26.8 points higher on Toolathlon than the release it replaced. GLM-5.3 gained 24.8 on the same board. Claude Opus 5.5 added 20.7 on SWE-bench Pro, and Claude Sonnet 5.5 added 18.1 there six days later. Scores via https://t.co/sxXugN92FJ, week of Sep 28.Opus 5.5X / TwitterMergeOct 1, 2026
RT @darylginn: Using @paper with Opus 5.5 a lot for design work lately. I don't think I can go back to doing things manually anymore. It's…Opus 5.5X / TwitterPaperOct 1, 2026
The best open-weight model for code is third overall, at a 23rd of the price of the model above it. MiMo-V2.6-Pro scores 60.9 on SciCode against Claude Opus 5.5's 66.9, and runs $0.87 per million output tokens against $20. Kimi K3 and GLM-5.3 follow within two points of it. Six points for $19 per million tokens is a real tradeoff, and it's the kind a Build-Your-Own-Router policy on Merge Gateway lets you set per request rather than once. Scores via https://t.co/TeTHwS52d4, week of Sep 28.Opus 5.5X / TwitterMergeSep 30, 2026
GPT-6.1 Sol promises similar performance to last week's GPT-6 release at one-fifth the token cost of Astra. Does it deliver? We re-ran our eval on how well models can guide users through multi-step problems. GPT-6.1 Sol reached 78.6% problem-solving accuracy at token prices far cheaper than Astra, Opus 5.5, or Fable 5.1. The most notable improvement was how the new model guided users through teaching interactions. Our "teaching quality metric" for GPT-6.1 Sol rose from 64.1% to 77.6%, compared to Astra's Opus 5.5X / TwitterBraintrustSep 30, 2026
Keeping up with AI is its own full-time job In the last month, we got: • Claude Fable 5.1 • GLM 5.3 • Gemini 3.8 Flash • GPT-6 Astra • Meta Muse Spark 1.3 • Claude for Small Business In the last week, we got: • Claude Opus 5.5 • GPT-6 Sol • GPT-6 Luna • Claude Sonnet 5.5 In the past 24 hours, we got: • GPT-6.1 Sol • OpenAI Dots • Muse for Small Business What'd we miss? By next week, this list will be outdated, but your workflows don't have to beOpus 5.5X / TwitterZapierSep 30, 2026
RT @wadefoster: GPT-6.1 Sol just launched. It matches Opus 5.5 on business workflows for about a third of the price. @Zapier's Automation…Opus 5.5X / TwitterZapierSep 29, 2026
RT @wadefoster: Sonnet 5.5 is here, less than a week after Opus 5.5. It’s cheaper and just took the top spot on our AutomationBench. At M…Opus 5.5X / TwitterZapierSep 28, 2026
Claude Sonnet 5.5 is live in Zapier! It took the top spot on AutomationBench, scoring 44.7% on Max effort vs. Opus 5.5's 40% Best for marketing, finance, and ops workflows Full scorecard with cost per task: https://t.co/tfc1e75dGxOpus 5.5X / TwitterZapierSep 28, 2026
Checking in on Opus 5.5 ~1 week after launch. It's the #1 model in share of spend and share of tokens among Anthropic models on OpenRouter Switching from Opus 5 has been particularly rapid https://t.co/K7kWAZQn21Opus 5.5X / TwitterOpenRouterSep 28, 2026
When does more reasoning stop being useful and start becoming overhead? We ran Open AI and Anthropic’s best models: GPT-6 Astra and Claude Opus 5.5 at every reasoning level on our new benchmark, MysteryMechanism. Thinking longer delivers accuracy gains with the same model at as much as 20x the cost. Inspecting traces, the models use their effort level in very different ways.Opus 5.5X / TwitterValsSep 28, 2026
Sonnet 5.5 beats Opus 5.5 on Terminal-Bench 4.0 at half the cost. It uses far fewer tokens than previous models to do the same work, showing costs reduced up to 30% per task. It also generates output 30% faster than its predecessor, making it the fastest Sonnet model to date. https://t.co/cLW1LDMwCMOpus 5.5X / TwitterClineSep 28, 2026
RT @HKrackDev: I had Sonnet 5.5 and Opus 5.5 build the Alpine Charlet head2head... Sonnet finishes 50% faster and the quality is just as…Opus 5.5X / TwitterCodeRabbitSep 28, 2026
Anthropic's "mid-tier" model is now the second best on the Vals Index. Sonnet 5.5 debuts at #2 at just 0.47 points behind Opus 5.5, and Anthropic holds the top 4 spots. https://t.co/kumflzZBWCOpus 5.5X / TwitterValsSep 28, 2026
Shipped. At lightspeed. 🚀 Heres the latest you can do in Replit: - Build apps for @Meta Quest VR - Create Replit apps through @Muse - Link payments with @airwallex - Build with new models: GPT-Sol 6, GPT-6 Luna, Opus 5.5 - Chart data and gain insights directly through chat - For Enterprise: Request and control access within workflows More at our changelog: https://t.co/COgha3boCaOpus 5.5X / TwitterReplitSep 28, 2026
The new #1 on SWE-bench Pro costs less than the model it passed. Claude Opus 5.5 debuted at 89.9% on September 22, 9.9 points clear of Claude Fable 5, and runs $20 per million output tokens against Fable 5's $50. SWE-bench Pro is the harder cut of SWE-bench. The average across all 59 models on it is 57.6% Scores via https://t.co/sxXugN92FJ, week of Sep 28.Opus 5.5X / TwitterMergeSep 28, 2026
RT @cvaneenige: Stream in 20 seconds, using Opus 5.5 @getstream_io https://t.co/d6iPyGHMoROpus 5.5X / TwitterStreamSep 28, 2026
RT @bentlegen: Opus 5.5 product vids aren't just good because they look great they're good because they can explain the product better tha…Opus 5.5X / TwitterModemSep 28, 2026
In this video, we break down what Claude Opus 5.5 is, what’s new, why its stronger agentic capabilities matter, and what happens as AI agents become more autonomous. Claude Opus 5.5 is an advanced artificial intelligence model released by Anthropic on September 22, 2026, designed specifically for long-running, multi-step agentic coding and complex knowledge work. 👉 Try Agent Testing Now: https://t.co/f2QAVJsULb As agents start handling longer workflows, calling tools, coordinating multiple steps, and makOpus 5.5X / TwitterTestMu AISep 28, 2026
RT @liran_tal: Opus 5.5 one-shotting a Snyk VulnBench reel (the securing AI benchmark work)... does not look as impressive as I hoped it t…Opus 5.5X / TwitterSnykSep 28, 2026
A faster spy trailer, done in about 30 minutes on Atlas Cloud 🎬 GPT Image 2.5 made the cast and locations. Seedance 2.5 generated 14 clips at 1080p, dialogue included. MiniMax Music 3.0 wrote the score. Opus 5.5 ran it all through one API key.Opus 5.5X / TwitterAtlas CloudSep 28, 2026
One brief, and Opus 5.5 took it all the way to a finished 60-second movie trailer on Atlas Cloud 🎬 GPT Image 2.5 built the characters and locations. Seedance 2.5 generated 7 clips at 1080p, voiceover included. MiniMax Music 3.0 wrote the score. One API key for all of it.Opus 5.5X / TwitterAtlas CloudSep 27, 2026
This Week in Replit 1) Three new models in Replit Agent: GPT-6 Sol, GPT-6 Luna Fast, and Claude Opus 5.5. Experiment with them and see how they fit your flow. 2) Muse can now make apps in Replit. Replit is an approved connector inside Muse, rolling out now. Go to Connectors and look for Replit. 3) Your next app could be in VR. Describe an app, Replit builds it for Meta devices. Announced this week at Meta Connect. → https://t.co/hGbdhDsmYwOpus 5.5X / TwitterReplitSep 26, 2026
It was a big week for model releases: Claude Opus 5.5, GPT-6 Sol, GPT-6 Luna, Grok 4.7, MiMo V2.6 Pro / Flash all dropped. We ran them on the Vals Index and across our other benchmarks . Here's the recap in case you missed it 🧵 https://t.co/5q5Svj88SuOpus 5.5X / TwitterValsSep 26, 2026
Moving the SonarQube Remediation Agent to @claudeai Opus 5.5 improved every quality measure we track and cut the cost of each fix by roughly 58%. 🖐️🎤 Details ⬇️ https://t.co/vP0PkLFJr1Opus 5.5X / TwitterSonarSep 25, 2026
RT @Thom_Wolf: ok what the fuck. this was a dry google doc a moment ago gave Opus 5.5 my Open Alignment brainstorm gdoc and asked for a vi…Opus 5.5X / TwitterHugging FaceSep 25, 2026
Opus 5.5 🤝 Brandfetch MCP Just one prompt... and - It wrote the entire script. - Found every logo used in the video using the MCP. - Designed the motion graphics itself. No stock assets. No manual design. No video model. We’re not ready for how fast this is getting. https://t.co/3j8yjTUKjOOpus 5.5X / TwitterBrandfetchSep 25, 2026
How does "thinking harder" change how models perform on math proofs? We ran GPT-6 Astra, Opus 5, and Opus 5.5 on every reasoning level: low, medium, high, xhigh, and max on our benchmark, Proof Bench v1.1. https://t.co/djytmJfUNxOpus 5.5X / TwitterValsSep 25, 2026
RT @amos_gyamfi: Opus 5.5 built me 184 SwiftUI AgentThinking text animations with Claude Code’s spinner verbs, from Accomplishing to Zigzag…Opus 5.5X / TwitterStreamSep 24, 2026
RT @kevinjosethomas: prime agent v0.9.6 is out: ◆ Support for GPT-6 Sol, Opus 5.5, and Grok 4.7 ◆ /mcp plugin catalog with one-click conne…Opus 5.5X / TwitterPrime IntellectSep 24, 2026
Claude Opus 5.5 is live in Model Garden in Gemini Enterprise https://t.co/zvbOjCDgVqOpus 5.5X / TwitterGoogle CloudSep 24, 2026
Have you tried Opus 5.5 yet with Paper? We are hearing good things about similar quality with reduced cost. @harcourttt takes it for a spin in this video -- what's your take on the newest model from Anthropic? https://t.co/rN79sMOt0LOpus 5.5X / TwitterPaperSep 24, 2026
RT @thisiskp_: How many mind blowing Claude Opus 5.5 demos have you watched so far? Well, now you can watch so many more all at once place…Opus 5.5X / TwitterNetlifySep 24, 2026
Concept art to 3D model to a finished Blender shot, with the agent calling every model itself. Claude Opus 5.5 and GPT-6 Sol both plugged into the Atlas Cloud MCP, generated concept art with GPT Image 2.5, turned it into 3D with Tripo, then built the scene in Blender. Opus ran 10 generations, Sol ran 2. Image models and 3D models, one MCP.Opus 5.5X / TwitterAtlas CloudSep 24, 2026
Claude Opus 5.5 is faster and cheaper than Fable 5.1, but 51 of its solves came from recalling fixes, not writing them. With recall removed: 33.5% SecPass, third on the Agent Security League. https://t.co/BFWaxkzM0P https://t.co/MvG4SLivYCOpus 5.5X / TwitterEndor LabsSep 24, 2026
RT @martinmitrevski: Opus 5.5 is pretty awesome! It helped me build a small paddling game based in Ohrid - Europe's most beautiful place ac…Opus 5.5X / TwitterStreamSep 24, 2026
RT @skalskip92: Opus 5.5 is the best "vision" model Anthropic ever made better than Fable 5, better than GPT-6 Sol, worse than GPT-6 Astra…Opus 5.5X / TwitterRoboflowSep 24, 2026
A video model and an image model, chained together by an agent. Claude Opus 5.5 and GPT-6 Sol each generated a parkour shot with Seedance 2.5, pulled the frames, restyled them with GPT Image 2.5, and cut it all back together. Nobody touched the edit in between. Both models run on one Atlas Cloud API key, so the agent never has to switch platforms.Opus 5.5X / TwitterAtlas CloudSep 24, 2026
Two frontier agents, one Atlas Cloud MCP. https://t.co/AyI4Rn8KIo Claude Opus 5.5 and GPT-6 Sol got the same brief, wrote their own prompts, and called GPT Image 2.5 themselves. One shot each. Opus hid a skeleton in the White Bone Demon's shadow. Sol opened a scripture with nothing written in it. Plug the MCP into your agent and it can run image, video and 3D models on Atlas Cloud.Opus 5.5X / TwitterAtlas CloudSep 24, 2026
Claude Opus 5.5 takes the #1 spot on Parallel's Search Capability Leaderboard, with a +4.3 gain over the second-highest score from Opus 5. https://t.co/z6xAEhhMHgOpus 5.5X / TwitterParallel Web SystemsSep 23, 2026
We put @claudeai Opus 5.5 to work on a $19 million reinsurance recovery where the right answer was buried across 14 documents in Box. In this Box Agent demo, Opus applies an amended occurrence window, reconciles claim payments with other recoveries, and separates estimated recovery from what can actually be called today. The records support $9.8 million in estimated recovery, but only a $3 million additional cash call. The agent generates a cited review for the insurer’s reinsurance team, with the outstanOpus 5.5X / TwitterBoxSep 23, 2026
RT @skalskip92: going live tomorrow to test GPT-6 Sol and Luna, Opus 5.5, MiMo-V2.6 Pro & Flash and Grok 4.7 Sep 24th, 17:00 CEST / 11:00…Opus 5.5X / TwitterRoboflowSep 23, 2026
/bro and i-have-adhd are OSS skills designed to make LLM writing more concise and conversational, much like Opus 5.5 claims. But can AI de-slop itself? /bro reduced writing quality by about 22–26 percentage points and task correctness by 17 points. i-have-adhd showed no reliable improvement on writing quality, while task correctness fell by about 11 points for Opus and 3 points for Fable. Results may differ on other tasks, but we found that style instructions still need to be evaled alongside task performOpus 5.5X / TwitterBraintrustSep 23, 2026
Claude Opus 5.5 is now available in the Warp Terminal and the Warp Agent CLI. https://t.co/vO3qRpHCYNOpus 5.5X / TwitterWarpSep 23, 2026
GPT-6 Luna, GPT-6 Sol, Claude Opus 5.5. Three more for the model picker. Available now in Software Factory: https://t.co/nksDX7nPLQOpus 5.5X / Twitter8090Sep 23, 2026
Anthropic released Claude Opus 5.5 today. 85% fewer sandbox escape attempts. Lowest prompt injection rate tested. 40% lower cost than Opus 5. But model-level security and enterprise runtime protection are two different things. What Opus 5.5 doesn’t include, and how enterprises close the gap → https://t.co/JjKptndoPO #ClaudeOpus55 #AISecurity #AgenticAI #CISOOpus 5.5X / TwitterNeuralTrustSep 23, 2026
Opus 5.5 is available in Agent Runners and AI Gateway, so the agent building your app and the app it builds can both run on the same model. Set the model string to claude-opus-5-5 and use the effort setting on the output config when you want it to work harder on a given request. https://t.co/Rg6feKFmyWOpus 5.5X / TwitterNetlifySep 23, 2026