Claude
Home All pages
All vikis
007 First LightAIAion 2AnantaARC RaidersArcheAge ChroniclesCall of Duty: Modern Warfare 4ChatGPTChrono OdysseyCivilization 3Civilization 4ClaudeCrimson DesertDeadlockDiablo 4Diablo 5EnshroudedFableFinal Fantasy ResonanceFootball Manager 27Forza Horizon 6God of WarGrand Theft Auto 6Heroes of Might and Magic 3
Start your own viki

Photos

Claude's Settings with Capabilities selected in the left menu. A card for Code execution and file creation is switched on. Below, a Skills section marked Preview, with an Upload skill button, lists acme-brand-guidelines (added today, on), acme-internal-comms (added yesterday, a cursor on its switch), and Anthropic's algorithmic-art, canvas-design and skill-creator, each switched on. A chat titled Acme Corporation Sales Call. The user asks Claude to review their Google calendar, emails and documents before a sales call with Acme Corporation tomorrow. Claude says it will gather information from Google Calendar, emails and documents; a step reads Searched recent events with 4 results, then Let me check your Gmail for correspondence with Acme Corporation above a step reading Searching messages, with a loading spinner. Research BETA is on and the model is Claude 3.7 Sonnet. A Claude window greeting Good afternoon, Alex above a prompt box reading How can I help you today?, with a plus button, a settings button and a Research button marked BETA, and the model menu set to Claude 3.7 Sonnet. Below, a From Drive card suggests prompts such as Show me patterns in how my documents are being used and Surface unexpected connections across my different documents, with a hand cursor on the second, on an orange background. A project page titled Customer Call Transcripts, described as holding the call transcripts of a prospective customer, Acme Inc., tagged Sales with three member avatars. Below a chat box, Mindy shared a chat quoting Acme's favorable reviews, titled Product Sentiment. On the right, Project knowledge lists three PDFs, Customer feedback transcripts, Survey results and NPS score, each added a few seconds ago, on a purple background. A Claude project about new hire onboarding. In the Project knowledge panel, a hand cursor rests on a dashed button, + Set custom instructions, and a dark tooltip reads Set instructions for how Claude should behave and respond in this project. Below are four files added today: [2024] Marketing team and [Q1 2024] All-hands meeting notes as DOCX, A guide to performance cycles and Marketing, Team charter as PDF. Under the heading Introducing Claude Managed Agents, a console page for a session, Build investment thesis for BuyCo, marked Idle, with chips for merges-and-acks, env, 1 file and production-vault, 22 hours ago, 5m 34s and 71.0k / 5.7k tokens. A colored timeline shows the orchestrator and three parallel agents; tabs read Orchestrator, Reporter, Analyst and Forecaster. The transcript lists Session start, a user Data Analysis Task on BuyCo retail orders, agent and Bash steps that unzip the dataset and run prepare.py, and Data is ready. Now I'll delegate to all three agents in parallel. A desktop window with a chat titled Pricing Comparison. The user asks to start a pricing comparison and to grab the latest pricing from Northwind's site; Claude replies that it will pull up the website, with steps Used Browser, Opening browser and Using browser. On the right, a browser tab at northwind.example/pricing shows One price. No surprises. and three plans: Starter $12/mo, Team $29/mo marked Popular, and Business $79/mo. A Claude chat titled Monthly new customer lands by segment. The user asks to pull up the quarterly sales revenue chart in Hex. Claude, through the Hex connector, thinks for 31 seconds, searches for tables named customer, lands, segment, date and created, finds a related Hex project, Sales Revenue Dashboard (endorsed, with a small stacked bar chart), and lists modeled data fields such as campaign_members, Mql cohort month, with a tooltip describing cm_cohort_month. Anthropic's Claude logo, an orange-red starburst of thin radiating strokes followed by the wordmark Claude in a dark serif typeface, centered on a plain cream background. A Slack thread in #checkout-incidents. Marcus Nguyen writes at 2:14 PM: @Claude checkout error rate jumped from 0.3% to 6% right after the 14:00 deploy, can you dig in and figure out what broke? Claude, marked AGENT, answers that it is tracing the spike, with checked steps: pulled the deploy diff from GitHub (#4821), queried Datadog with 502s isolated to cart-service, correlated a 5x spike in DB connection-pool timeouts, compared error rates by region, and is writing a root-cause summary and rollback recommendation. A link reads Open session in Claude. A Slack window on #eng-tickets. Nathan Wilson reports at 9:02 AM that auth is throwing 502s on token refresh since the deploy, likely a race condition in services/auth. Elena Barna writes @Claude fix the bug in this thread and open a draft PR. At 9:06 AM Claude, marked AGENT, explains a race in services/auth/token_cache.go exposed by deploy a3f9c1e dropping a per-request mutex, and posts Draft PR #48217, CI green, +71 -12, with a link to Open session in Claude. The words Claude Sonnet 5.5 in white serif type over a view of Earth's curve, blue ocean and white clouds, seen through a square window in a dark spacecraft interior. The Claude Science beta app in a Literature review session. The user asks for a review of cross-species single-cell RNA-seq integration as a LaTeX document and compiled PDF. Claude dispatches five parallel retrieval tracks (PubMed, bioRxiv, OpenAlex, CELLxGENE and orthology-free embedding methods); a Reviewer card shows a Warn finding that one PubMed ID was assigned to two methods, which Claude acknowledges and corrects. On the right, review.pdf shows the paper's title page with a citation chart and two-column text. The Claude Science beta app with a sidebar of sessions such as Samosa Manuscript, Samosa Editorial Review and Samosa Model Training. The middle pane shows samosa_fig1a.png, a scatter plot of 138 species and 5,672 cell types colored as neuron, muscle and immune, with a comment box reading these labels are hard to see and a Send button. The right pane, version 2, has tabs Code, Execution Log, Messages, Environment and Review, a Download script button, input CSV files and the matplotlib code. The words Claude Opus 5.5 in white serif type across a dark sky with an orange and blue horizon glow, in a central panel flanked by collage strips of yellow paper and textured orange gills. A table with columns Claude Mythos 5 / Fable 5 (highlighted), Claude Mythos Preview, Claude Opus 4.8, GPT 5.5 and Gemini 3.1 Pro. Rows include SWE-Bench Pro (80.3%, 77.8%, 69.2%, 58.6%, 54.2%), FrontierCode Diamond, GDPval-AA (1932 for Mythos 5 / Fable 5), GDP.pdf, Blueprint-Bench 2, AutomationBench, OSWorld-Verified (85.0%, with Mythos Preview's 85.4% highlighted), Legal Agent Benchmark, Humanity's Last Exam, BioMysteryBench, Terminal-Bench 2.1 (88.0%), ExploitBench (78.0%) and HealthBench Professional (66.0%). A methodology note says the two models score within 1 to 3 points and the table shows the higher score. A table with columns Sonnet 5 (highlighted), Sonnet 4.6 and Opus 4.8 For reference. SWE-bench Pro: 63.2%, 58.1%, 69.2%. Terminal-Bench 2.1: 80.4%, 67.0%, 82.7%. Humanity's Last Exam: 43.2%, 34.6%, 49.8% with no tools and 57.4%, 46.8%, 57.9% with tools. OSWorld-Verified: 81.2%, 78.5%, 83.4%. GDPval-AA v2: 1618, 1395, 1615. A Chrome window on zillow.com showing Seattle WA real estate for sale: a map dotted with red price tags from 559K to 750K and listing cards for $749,950 and $663,000 homes. A Stop Claude button sits at the foot of the page. In the Claude side panel, titled Seattle home search, the user asks for a 3-bedroom house in Seattle under $800K with a garage and at least 1500 sqft, top 5 options, and Claude shows Running command..., with Opus 5 High as the model. A Chrome window on calendar.google.com in week view for Aug 18 to 22, 2026, with a meeting card for Darshan / Si Min on Monday, Aug 18, 1:00 to 1:30pm open over the grid and a Stop Claude button at the bottom. The Claude side panel, titled Calendar review, holds the request to go through this week's meetings that need rooms, check room availability and book spaces by meeting size and location, and shows Running command..., with Opus 5 High as the model. The words Claude Haiku 5.5 in black serif type on an off-white background, set along a thin curved line that runs through three collage blocks: yellow paper on the left, a crossed red and black pattern in the middle and blue paper on the right. A bar chart titled Agentic coding, success rate in percent. On SWE-Bench Pro, Claude Fable (orange) scores 80.3, Claude Opus 4.8 (green) 69.2 and GPT 5.5 (gray) 58.6. On FrontierCode, Claude Fable scores 29.3, Claude Opus 4.8 13.4 and GPT 5.5 5.7. The Claude Duet product logo, a rounded white tile holding an orange circle crossed by a looping black line, followed by the words Claude Duet in black serif type on a cream background. A design canvas showing a dark globe of North America with teal arcs linking cities such as Vancouver, Seattle, San Francisco, Los Angeles, Toronto, New York, Atlanta and Miami. A toolbar has a Tweaks switch turned on, Comment and Edit buttons. A dark side panel offers Theme (Dark or Light), Breakpoint (Desktop, Tablet, Mobile) and Network settings: arc color, arc width, arc glow, arc density, city size and pulse speed sliders, and checkboxes for Show arcs, Show cities and City labels. A dark terminal window headed Claude Code v2.0.0. A dashed orange box says Welcome back Meaghan! above an orange pixel-art creature and the line Sonnet 4.5, Max 20x, /users/meaghan/code/apps. To the right, Recent activity lists updates from 1 minute to 1 week ago, and What's new lists /agents to create subagents, /security-review for review agent and ctrl+b to background bashes. A prompt line at the bottom suggests: try edit <filepath> to ... A black terminal running Claude Code v2.0.12 with Sonnet 4.5 in /Users/daisy/code. After the typed command /plugin, a bordered panel titled claude-code-marketplace, Manage Plugins lists three enabled plugins: code-review (Commands for code review workflows, v1.0.0), typescript-lsp (Plugin from typescript-lsp@claude-code-marketplace) and security-guidance (Security reminder hook, v1.0.0). Hints below read Space to toggle enable, u to mark update, Delete to mark uninstall, Enter for details, Esc to back. A light panel titled Claude Code with an empty prompt box reading Ready to code..., a GitHub repository selector set to acme/production/apps, and an open dropdown with a Search branches field above the branches main, nate/module-ui-fix, meaghan/4o94u390-bugs, george/version-update, susan/onboarding-review-task and nate/gui-observer, on a blue background. A dark editor side panel. The user asks: can you add tests for this file? with localizationUtils.ts attached. Claude answers Sure, let me take a look at the file, then shows the steps Read localizationUtils.ts (Read 256 lines) and Write src/test/localization.ts with green highlighted lines importing formatDate, translateKey and getCurrency and describing LocalizationUtil tests. An orange spark says Pondering..., and the input box below reads Queue another message... with the mode Ask before editing. A line chart titled Agentic coding by effort level, Artificial Analysis Coding Agent Index. The index score in percent runs up the side from 45 to 70 and cost per task in US dollars along the bottom on a log scale from $1.50 to $15.00, with points for the effort ladder low, medium, high, xhigh and max. Opus 5 (orange) rises from about 57 at $2.20 to about 67 near $8; Fable 5 (yellow) runs from about 57 to 66 at higher cost; Opus 4.8 (blue) from about 47 to 61; GPT-5.6 Sol (light gray) from about 54 to 67. A bar chart titled Software engineering, SWE-bench verified, accuracy in percent. Opus 4 scores 72.5% (79.4% with parallel test-time compute), Sonnet 4 72.7% (80.2%), Sonnet 3.7 62.3% (70.3%), OpenAI Codex-1 72.1%, OpenAI o3 69.1%, OpenAI GPT-4.1 54.6% and Gemini 2.5 Pro Preview (05-06) 63.2%. A table with columns Claude 3 Opus, Claude 3 Sonnet and Claude 3 Haiku (outlined in orange), GPT-4, GPT-3.5, Gemini 1.0 Ultra and Gemini 1.0 Pro. Opus scores 86.8% on MMLU, 50.4% on GPQA Diamond, 95.0% on GSM8K, 60.1% on MATH, 90.7% on MGSM, 84.9% on HumanEval, 83.1 on DROP, 86.8% on BIG-Bench-Hard, 96.4% on ARC-Challenge and 95.4% on HellaSwag, its scores in green. A table with columns Claude 3.5 Sonnet (new) and Claude 3.5 Haiku outlined in orange, then Claude 3.5 Sonnet, GPT-4o, GPT-4o mini, Gemini 1.5 Pro and Gemini 1.5 Flash. The new 3.5 Sonnet scores 65.0% on GPQA Diamond, 78.0% on MMLU Pro, 93.7% on HumanEval, 78.3% on MATH, 16.0% on AIME 2024, 70.4% on MMMU, 49.0% on SWE-bench Verified (against 33.4% for the earlier 3.5 Sonnet) and 69.2% retail and 46.0% airline on TAU-bench. A footnote says the tables exclude OpenAI's o1 models. A chart with Intelligence (benchmark scores) up the side and Cost (price per million tokens) along the bottom. A dashed curve rises from Claude 3 Haiku through a faded Claude 3 Sonnet to Claude 3 Opus; an arrow rises straight up from Claude 3 Sonnet to Claude 3.5 Sonnet, which sits higher than Opus at Sonnet's cost. The Claude app in a project called Project Operations. A user who attached a Data TXT file asks Claude to build a scatter plot of page load time in seconds against user session duration in minutes; Claude replies with a React component card, Scatter Plot for User Engagement Metrics. On the right a panel with Preview and Code tabs shows the code: imports from recharts, a data array of pageLoadTime and sessionDuration pairs, and a ResponsiveContainer with a ScatterChart. The model menu reads Claude 3.5 Sonnet. The word ANTHROPIC in heavy black geometric capitals, with the I drawn as a backslash, centered on a plain cream background.

Hotkeys

/
Jump to the search box
? or Ctrl + /
Open this list
Esc
Close this list, the photo viewer, or the account menu

Photo viewer

← → or A D
Previous or next photo
↑ ↓ or W S
Previous or next photo
Home End
First or last photo
I
Show or hide the photo's info
Z
Clickless on or off: a pointer over a thumbnail shows it
N
Compare with the last photo: back to the one shown before this, and N again comes back
Q
Move the thumbnails to the next edge: top, bottom, left and round
, or Ctrl + ,
Open or close the viewer's settings
V
Clickless pan: a zoomed photo follows the pointer, rather than being dragged
Shift
Browse on or off
B
Show or hide the button bar
T
Show or hide the thumbnails
1
100% zoom on or off
0
Rate the photo as no opinion; again takes the answer back
+ -
Rate the photo a step higher or lower
F
Full screen: the photo alone. F again puts everything back.