SYLIN

Anthropic

Maker · 18 entries

Everything from Anthropic

NameWhat to knowPrice or memoryRanks
Claude FableModelstrong Top-ranked on most independent leaderboards, but expensive and beaten on some tasks.$8 / $40#1 agents#1 coding#1 documents
Claude OpusModelstrong Still a top-three model for agents and coding, now behind Fable 5.2.#2 documents#2 image understanding#2 web development
Claude black-hole physics resultsPaperunproven Anthropic's claim of two solved physics problems is unreviewed and cannot yet be checked.
Claude CodeAgentpromising Widely used coding harness; praise is real but mostly from its maker and informal users
Claude HaikuModelmixed Cheap and capable, but independent tests rank it behind Kimi K3 and GLM 5.3
Claude MeetingsApp or serviceunproven Beta Google Meet assistant for Team and Enterprise plans; little outside testing, privacy worries
Claude Code ProjectsApp or serviceunproven New Anthropic multi-agent coding workspace in limited beta; only the maker and a few commenters have spoken
Claude DuetApp or serviceunproven Anthropic's always-on assistant is rolling out in the US, but no one has tested it yet
Anthropic safety-evaluation recommendationsPaperAnthropic published 10 recommendations for designing, running and reporting safety evaluations
Claude Managed AgentsApp or serviceSelf-hosted private beta: run harness, sandboxes and session history in your own cloud or on-prem; adds agent identity, OTel, kill switches, evals
Claude SonnetModelmixed Near the top in intelligence, but costly and beaten on value by cheaper rivals
Claude MCP AppsApp or serviceExpanded support for app tools and embedded sites in public preview; apps portable to other clients supporting the extension
Claude MythosModelpromising Mythos 5.2 looks stronger than 5.1 on security tasks, per vendor-side numbers only
Claude Whiteboards, Sheets and DataApp or serviceClaude Whiteboards and Sheets rolling out; Claude Data now generally available
Claude enzyme-system discoveryPaper~950 Claude agents scanned DNA for reverse transcriptases and flagged a possible CRISPR-like system; function untested and contested
Anthropic Fermat's Last Theorem formalizationPaperClaude agents formalized Fermat's Last Theorem in Lean in about 11 days: roughly 13M lines, 29,500 theorems used in the proof
Anthropic Python SDKFrameworkv1.0.0 made the same switch to HTTPX2, weeks after OpenAI's SDK
Model Hardware StandardFrameworkResearch preview of a shared spec letting AI agents operate lab and manufacturing devices via MCP, CLI or code; built with HHMI Janelia

Plans

What each one gives you.
Claude Pro $20/month billed monthly, or $17/month billed annually1x baseline usage per 5 hours and per weekClaude Opus, Claude Sonnet, Claude Haiku, Claude Fable
Claude Max $100/month5x Pro usage per 5 hours and per week; $100 monthly API creditsClaude Opus, Claude Sonnet, Claude Haiku, Claude Fable
Claude Ultra $200/monthFable up to half the weekly limitClaude Opus, Claude Sonnet, Claude Haiku, Claude Fable