See how people build with AI.

Discover what builders use, compare AI models, and follow the companies and capital behind them.

Explore leaderboardsTrack your usage

Humans

All humans
  1. 1YeonGyu-Kim (@code-yeongyu)Top model and agent: Claude Opus 5.5, Senpi (OmO Native)Tokens: 882.6BEstimated cost: $430.6K
  2. 2Jeff Emanuel (@Dicklesworthstone)Top model and agent: Claude Opus 5.5, Claude CodeTokens: 593.2BEstimated cost: $206.5K
  3. 3ualtinokTop model and agent: Claude Opus 5.5, JunieTokens: 581.4BEstimated cost: $135.9K
  4. 4Shun Kakinoki (@shunkakinoki)Top model and agent: GPT-6 Luna, Codex CLITokens: 529.6BEstimated cost: $36.5K
  5. 5registered2ndTop model and agent: Claude Opus 5.5, Claude CodeTokens: 347BEstimated cost: $85.7K

Models

All models
  1. 1Claude Opus 5.5 by AnthropicActive users: 711Change in active users against the previous period: −103Estimated cost: $4.8M
  2. 2GPT-6.1 Sol by OpenAIActive users: 669Change in active users against the previous period: −57Estimated cost: $878.8K
  3. 3GPT-6 Astra by OpenAIActive users: 561Change in active users against the previous period: −249Estimated cost: $1.9M
  4. 4Claude Sonnet 5.5 by AnthropicActive users: 554Change in active users against the previous period: 0Estimated cost: $230.7K
  5. 5Claude Haiku 5.5 by AnthropicActive users: 497Change in active users against the previous period: +170Estimated cost: $11K

Trusted by Professionals at

  • Microsoft
  • Amazon
  • Google
  • ByteDance
  • Tencent
  • Alibaba
  • Intel
  • Toss
  • Hashed

Humans

Find people building with the tools you use.

Every public profile shows which models and agents a person uses, how many tokens they run, and how often they work with AI. Profiles come from usage people chose to submit.

Browse the humans leaderboard

Jeff Emanuel #2 by tokens, last 30 days

@Dicklesworthstone

Usage · Last 30 days (Oct 18 – Nov 16, 2026)

Tokens
593.2B
Est. cost
$206.5K
Active days
13 of 30

Stack, by tokens in the period

Models
  • Claude Opus 5.5Anthropic
  • GPT-6.1 SolOpenAI
  • Claude Sonnet 5.5Anthropic
Agents
  • Claude Code
  • Codex CLI
  • Oh My Pi

Models

Compare models before you build.

See how many people use a model, which agents they use it through, and how many tokens it handles.

Compare models

Claude Opus 5.5 #1 by active users, last 30 days

Anthropic

Active usersPeople who used it in the period
711−103 vs previous period
AdoptionShare of all 949 active accounts
75%
TokensInput, output, cache, and reasoning tokens
12.4T
Est. costAt public list prices, not what people paid
$4.8M
Top agentsShare of this model's tokens
  • Claude Code79%
  • Senpi (OmO Native)10%
  • Oh My Pi4%

GPT-6.1 Sol #2 by active users, last 30 days

OpenAI

Active usersPeople who used it in the period
669−57 vs previous period
AdoptionShare of all 949 active accounts
70%
TokensInput, output, cache, and reasoning tokens
3.6T
Est. costAt public list prices, not what people paid
$878.8K
Top agentsShare of this model's tokens
  • Codex CLI79%
  • Claude Code5%
  • Oh My Pi4%

Capital

The money behind the models you use.

Funding rounds, compute contracts and capital spending across AI labs, clouds and chips. Every number comes from a filing or a company announcement.

Explore capital

Equity Raised · 2026 So Far

$237B

4 closed AI-lab rounds · plus €3B in EUR, not converted

Rounds

Compute Deals · 2026

12 deals

Largest stated value ≥$121.3B, Anthropic ↔ Alphabet

Compute

Cash Capex · Q3 2026

$223.8B

9 companies, cash paid for property and equipment

Spending

Largest Valuation

$965B

Anthropic, post-money, May 28, 2026

Anthropic Series H

These are four different money types. They are never added together. How we count

Labs behind the models you use

All companies
  1. AnthropicTop model: Claude Opus 5.5Last round: Series H$965B post · May 2026Largest compute deal: ≥$121.3Bwith Alphabet
  2. OpenAITop model: GPT-6.1 SolLast round: Equity round$852B post · Mar 2026Largest compute deal: >$20Bwith Cerebras
  3. DeepSeekTop model: DeepSeek V4.1 FlashLast round: PrivateLargest compute deal: None tracked
  4. Z.aiTop model: GLM-5.3 FlashLast round: Public · Listed in Hong KongLargest compute deal: None tracked
  5. AlphabetTop model: Gemini 3.8 Flash NLast round: Public · NASDAQ: GOOGLLargest compute deal: ≥$121.3Bwith Anthropic

Tokscale tokens: share of tokens from identified models among 949 Tokscale accounts, Oct 18 – Nov 16, 2026. Not market share. Valuations and contracts are company-stated or from filings.

Latest moves

All rounds
  1. Nov 10, 2026ComputeAnthropic discloses amended Google Cloud compute commitmentAnthropic ↔ Alphabet≥$121.3BS-1/A ↗
  2. Nov 10, 2026IPOAnthropic sets terms for its initial public offeringAnthropic—S-1/A ↗
  3. Nov 10, 2026InvestmentAnthropic $26 billion strategic private placement alongside its IPOAnthropic$26BS-1/A ↗
  4. Nov 9, 2026DebtAkamai prices two convertible-note seriesAkamai$2.2B8-K ↗

Instagram

Tokscale data, in your feed.

Cards on how people build with AI, drawn from Tokscale data, and on the money behind AI.

Follow @tokscaleai

Track your usage. Join the community.

Tokscale reads the session logs your coding agents already keep, on your machine. Nothing leaves it until you submit.

bunx tokscale@latest
Start tracking