TokenMaxxer
Live rankings

AI Model Rankings

Share

Live rankings from real developer usage. Models are ranked by the tokens developers actually run through Claude Code, Codex, Cursor and every other tool that syncs to TokenMaxxer. View all models

Usage data through Nov 19, 2026

Tokens tracked
1.13T
43.16B this week
Spend at list price
$661K
$12.31K this week
Models in use
150
from 11 providers
Tools
11
CLIs, editors, agents

Top Models

Weekly token usage of models across TokenMaxxer, last six months.

0
25B
50B
75B
100B
125B
  1. May 22 – 28, 2026. Total 4.6B. GPT-5.5 1.1B, Opus 4.8 169M, Others 3.3B.
  2. May 29 – Jun 4, 2026. Total 9.5B. Opus 4.8 4.5B, GPT-5.5 3.7B, Others 1.3B.
  3. Jun 5 – 11, 2026. Total 12B. GPT-5.5 4.8B, Opus 4.8 3.9B, Fable 5 1.8B, Others 1.5B.
  4. Jun 12 – 18, 2026. Total 10.3B. Opus 4.8 4.8B, GPT-5.5 2.7B, Fable 5 678M, Others 2B.
  5. Jun 19 – 25, 2026. Total 9.5B. Opus 4.8 6.4B, GPT-5.5 1.7B, Others 1.3B.
  6. Jun 26 – Jul 2, 2026. Total 14.9B. Opus 4.8 9.8B, Fable 5 1.9B, GPT-5.5 1.7B, Others 1.4B.
  7. Jul 3 – 9, 2026. Total 18.7B. Opus 4.8 9.8B, Fable 5 3.8B, GPT-5.5 1.1B, GPT-5.6 Sol 687M, Others 3.2B.
  8. Jul 10 – 16, 2026. Total 30.7B. GPT-5.6 Sol 11.6B, Opus 4.8 8.7B, Fable 5 4.8B, GPT-5.5 1.7B, Others 3.9B.
  9. Jul 17 – 23, 2026. Total 62.4B. GPT-5.6 Sol 36.1B, Opus 4.8 10.5B, Fable 5 6.1B, GPT-5.5 6B, Others 3.8B.
  10. Jul 24 – 30, 2026. Total 41.4B. GPT-5.6 Sol 15.3B, Opus 5 13B, Opus 4.8 4.6B, GPT-5.5 1.9B, Fable 5 1.8B, Others 4.7B.
  11. Jul 31 – Aug 6, 2026. Total 35.1B. GPT-5.6 Sol 14.1B, Opus 5 13.8B, Fable 5 1.9B, Opus 4.8 892M, GPT-5.5 26.8M, Others 4.4B.
  12. Aug 7 – 13, 2026. Total 40.9B. GPT-5.6 Sol 17.9B, Opus 5 12.9B, Fable 5 3B, Opus 4.8 986M, GPT-5.5 2.1M, Others 6.1B.
  13. Aug 14 – 20, 2026. Total 34.3B. Opus 5 20.9B, GPT-5.6 Sol 7B, Fable 5 2.7B, Opus 4.8 134M, GPT-5.5 21.4M, Others 3.5B.
  14. Aug 21 – 27, 2026. Total 36.9B. GPT-5.6 Sol 17B, Opus 5 15.4B, Fable 5 1.4B, Opus 4.8 185M, GPT-5.5 15M, Others 2.9B.
  15. Aug 28 – Sep 3, 2026. Total 75.9B. GPT-5.6 Sol 47.2B, Opus 5 13.4B, Fable 5 2.2B, Opus 4.8 244M, GPT-5.5 52.2M, Others 12.8B.
  16. Sep 4 – 10, 2026. Total 118B. GPT-5.6 Sol 71B, Opus 5 20.2B, GPT-6 Astra 16.9B, Opus 4.8 38.8M, GPT-5.5 11.9M, Others 9.9B.
  17. Sep 11 – 17, 2026. Total 22.8B. Opus 5 11.4B, GPT-6 Astra 6.5B, GPT-5.6 Sol 746M, Opus 4.8 37.2M, GPT-5.5 9.6M, Others 4.1B.
  18. Sep 18 – 24, 2026. Total 43B. Opus 5 22.2B, Opus 5.5 8.7B, GPT-6 Astra 4.9B, GPT-5.6 Sol 996M, Opus 4.8 202M, Fable 5 40.3M, Others 6.1B.
  19. Sep 25 – Oct 1, 2026. Total 47.3B. Opus 5.5 31.3B, GPT-6 Astra 4.9B, GPT 6 1 SOL 2.8B, Opus 5 1.8B, GPT-5.6 Sol 193M, Fable 5 20.9M, Opus 4.8 10.2M, Others 6.3B.
  20. Oct 2 – 8, 2026. Total 58.5B. Opus 5.5 33.3B, GPT 6 1 SOL 10.2B, GPT-6 Astra 9.8B, Opus 5 1B, Opus 4.8 770M, GPT-5.6 Sol 118M, GPT-5.5 165K, Others 3.3B.
  21. Oct 9 – 15, 2026. Total 44.6B. Opus 5.5 24.6B, GPT 6 1 SOL 7.3B, GPT-6 Astra 5.7B, Opus 4.8 304M, Opus 5 246M, GPT-5.6 Sol 6M, Others 6.4B.
  22. Oct 16 – 22, 2026. Total 66.6B. Opus 5.5 39.8B, GPT-6 Astra 11.8B, GPT 6 1 SOL 11.3B, Opus 4.8 429M, Opus 5 65.2M, Fable 5 14.9M, GPT-5.6 Sol 780K, Others 3.1B.
  23. Oct 23 – 29, 2026. Total 89.9B. Opus 5.5 48.4B, GPT 6 1 SOL 25.2B, GPT-6 Astra 9.5B, Opus 4.8 1.8B, Opus 5 177M, GPT-5.6 Sol 44.3M, Fable 5 6.8M, Others 4.7B.
  24. Oct 30 – Nov 5, 2026. Total 55.5B. Opus 5.5 27.6B, GPT 6 1 SOL 21.6B, GPT-6 Astra 1.3B, Opus 4.8 416M, Opus 5 165M, GPT-5.6 Sol 68.2M, Fable 5 324K, Others 4.2B.
  25. Nov 6 – 12, 2026. Total 51.4B. Opus 5.5 23.8B, GPT 6 1 SOL 12.1B, Opus 4.8 845M, GPT-6 Astra 632M, GPT-5.6 Sol 44.3M, Opus 5 23.6M, Others 13.9B.
  26. Nov 13 – 19, 2026. Total 43.2B. Opus 5.5 27.7B, GPT 6 1 SOL 8.7B, Opus 4.8 871M, GPT-6 Astra 625M, Opus 5 93.2M, GPT-5.6 Sol 18.8M, Others 5.2B.
May ’26Jul 31Oct 9
Share of the last 26 weeks

Each column is a trailing seven-day window; the last one ends today, so it matches every “this week” figure on the page. The chart opens on six months because most tools only started syncing in mid-2026, so earlier weeks mostly show the tools that keep long histories.

LLM Leaderboard

Models ranked by tokens processed across every tool on TokenMaxxer.

  1. 1.Opus 5.5by Anthropic
    27.7B tokens
    up 16%
  2. 2.GPT 6 1 SOLby OpenAI
    8.7B tokens
    down 29%
  3. 3.Claude Fable 5 2by Anthropic
    1.6B tokens
    down 25%
  4. 4.Claude Haiku 5 5by Anthropic
    1B tokens
    up >999%
  5. 5.Gemini 3 8 Flash Nby Google
    961M tokens
    down 2.8%
  1. 6.Opus 4.8by Anthropic
    871M tokens
    up 3.1%
  2. 7.GPT-6 Astraby OpenAI
    625M tokens
    down 1.2%
  3. 8.Fable 5.1by Anthropic
    445M tokens
    up 6.0%
  4. 9.Codex Auto Reviewby OpenAI
    406M tokens
    up 18%
  5. 10.Claude Sonnet 5 5by Anthropic
    307M tokens
    up 8.1%

Trailing 7 days (Nov 13 – Nov 19) against the 7 days before them.

Top Models by Tool

Each tool's leading models, sized by share of tokens on TokenMaxxer. Pick a tool to rank what runs inside it.

Claude Code48% of all tokens all time · 545B tokens
  1. 1Opus 5.5by Anthropic
    49%
    265B
  2. 2Opus 5by Anthropic
    27%
    145B
  3. 3Opus 4.8by Anthropic
    10%
    55.8B
  4. 4Fable 5by Anthropic
    5.2%
    28.5B
  5. 5Sonnet 5by Anthropic
    3.1%
    16.8B
  1. 6Fable 5.1by Anthropic
    2.3%
    12.4B
  2. 7Claude Haiku 5 5by Anthropic
    1.3%
    7B
  3. 8Claude Sonnet 5 5by Anthropic
    1.2%
    6.4B
  4. 9Claude Fable 5 2by Anthropic
    0.7%
    3.8B
  5. 10Sonnet 4.6by Anthropic
    0.3%
    1.6B

Cost per Session

Crunching individual sessions: what one coding-agent session typically costs, by session length, across finished sessions started in the last 60 days.

Cursor and Grok Build are left out: they report each request or prompt with its agent steps rolled into one row, so their sessions can't be grouped by how many model calls they ran.

Session Length

Crunching individual sessions: how long a coding-agent session runs, counted in model calls, across finished sessions started in the last 60 days.

A turn is one model call (a reply or a single agent step), not one message you typed, so one prompt can run dozens. Length is the time from a session's first turn to its last, idle gaps included. A session with a turn in the last 6hours is left out until it settles, so none is counted half-finished. Cursor and Grok Build are left out: they report each request or prompt with its agent steps rolled into one row, so their turn counts aren't comparable.

Market Share

Compare token share by coding tool on TokenMaxxer, week by week.

0%
25%
50%
75%
100%
  1. May 22 – 28, 2026. Total 4.6B. Cursor 77%, Codex 19%, Claude Code 3.7%, Antigravity 0.3%, OpenCode <0.1%, Pi <0.1%.
  2. May 29 – Jun 4, 2026. Total 9.5B. Cursor 57%, Codex 34%, Claude Code 7.7%, Antigravity 0.6%, OpenCode 0.1%, Pi <0.1%.
  3. Jun 5 – 11, 2026. Total 12B. Cursor 38%, Codex 38%, Claude Code 24%, Antigravity 0.6%, OpenCode <0.1%, Pi <0.1%.
  4. Jun 12 – 18, 2026. Total 10.3B. Claude Code 42%, Cursor 32%, Codex 22%, Antigravity 3.3%, OpenCode 1.0%, Pi <0.1%.
  5. Jun 19 – 25, 2026. Total 9.5B. Claude Code 44%, Cursor 36%, Codex 17%, Antigravity 3.1%, Pi <0.1%.
  6. Jun 26 – Jul 2, 2026. Total 14.9B. Claude Code 67%, Cursor 21%, Codex 12%, Antigravity 0.5%, Pi <0.1%.
  7. Jul 3 – 9, 2026. Total 18.7B. Claude Code 64%, Cursor 29%, Codex 5.9%, Antigravity 0.8%, Pi 0.1%.
  8. Jul 10 – 16, 2026. Total 30.7B. Claude Code 45%, Codex 31%, Cursor 20%, OpenCode 2.2%, Pi 1.2%, Antigravity <0.1%.
  9. Jul 17 – 23, 2026. Total 62.4B. Codex 58%, Claude Code 28%, Cursor 9.3%, Pi 4.3%, OpenCode 0.5%, Antigravity <0.1%, Others <0.1%.
  10. Jul 24 – 30, 2026. Total 41.4B. Claude Code 47%, Codex 41%, Pi 6.0%, Cursor 5.7%, Antigravity 0.2%, OpenCode <0.1%, Others <0.1%.
  11. Jul 31 – Aug 6, 2026. Total 35.1B. Claude Code 51%, Codex 28%, Pi 15%, Cursor 6.3%, Antigravity 0.7%, OpenCode <0.1%.
  12. Aug 7 – 13, 2026. Total 40.9B. Claude Code 51%, Codex 35%, Pi 12%, Cursor 2.2%, Antigravity 0.2%, OpenCode <0.1%.
  13. Aug 14 – 20, 2026. Total 34.3B. Claude Code 74%, Codex 17%, Pi 4.5%, Cursor 4.3%, Antigravity <0.1%, OpenCode <0.1%.
  14. Aug 21 – 27, 2026. Total 36.9B. Codex 49%, Claude Code 47%, Cursor 3.5%, OpenCode 0.1%, Antigravity <0.1%.
  15. Aug 28 – Sep 3, 2026. Total 75.9B. Codex 76%, Claude Code 22%, Cursor 1.7%, Antigravity 0.3%.
  16. Sep 4 – 10, 2026. Total 118B. Codex 76%, Claude Code 22%, Cursor 1.1%, Antigravity 0.2%, OpenCode <0.1%.
  17. Sep 11 – 17, 2026. Total 22.8B. Claude Code 58%, Codex 34%, Cursor 7.2%, OpenCode 1.0%, Antigravity <0.1%.
  18. Sep 18 – 24, 2026. Total 43B. Claude Code 80%, Codex 15%, Cursor 4.3%, OpenCode 0.7%, Antigravity <0.1%.
  19. Sep 25 – Oct 1, 2026. Total 47.3B. Claude Code 74%, Codex 24%, Antigravity 1.4%, Cursor 0.4%, OpenCode 0.1%.
  20. Oct 2 – 8, 2026. Total 58.5B. Claude Code 63%, Codex 35%, Cursor 0.8%, Antigravity 0.6%, OpenCode 0.3%, Qwen CLI <0.1%.
  21. Oct 9 – 15, 2026. Total 44.6B. Claude Code 69%, Codex 30%, Antigravity 1.1%, Cursor 0.3%, OpenCode 0.2%.
  22. Oct 16 – 22, 2026. Total 66.6B. Claude Code 63%, Codex 35%, Antigravity 0.9%, Cursor 0.4%, Droid <0.1%.
  23. Oct 23 – 29, 2026. Total 89.9B. Claude Code 59%, Codex 40%, Antigravity 1.0%, Cursor 0.2%, OpenCode <0.1%.
  24. Oct 30 – Nov 5, 2026. Total 55.5B. Claude Code 56%, Codex 42%, Antigravity 1.3%, Cursor 0.4%, OpenCode 0.3%, Qwen CLI <0.1%, Others <0.1%.
  25. Nov 6 – 12, 2026. Total 51.4B. Claude Code 54%, Codex 43%, Antigravity 2.2%, OpenCode 0.6%, Cursor 0.5%, Qwen CLI <0.1%, Others <0.1%.
  26. Nov 13 – 19, 2026. Total 43.2B. Claude Code 74%, Codex 23%, Antigravity 2.7%, OpenCode 0.3%, Cursor 0.2%, Others <0.1%.
May ’26Jul 31Oct 9
Share of the last 26 weeks
This week · Nov 13 – 19, 2026
  1. 1.Claude Code24 developers this week
    74%
    up +20.6 points
  2. 2.Codex13 developers this week
    23%
    down −20.5 points
  3. 3.Antigravity3 developers this week
    2.7%
    up +0.6 points
  4. 4.OpenCode1 developer this week
    0.3%
    down −0.3 points
  1. 5.Cursor1 developer this week
    0.2%
    down −0.3 points
  2. 6.Qwen CLI0 developers this week
    0.0%
    0
  3. 7.Others1 more tool
    <0.1%
    0

Shares are for the trailing 7 days (the chart's last column), with the change in share points against the 7 days before. Tools outside the chart's top 8 fold into Others. Cursor’s editor and cloud usage count as one tool, as do Antigravity’s app and CLI.

Price vs. Usage

What a million tokens costs at list price against how widely each model is used, over the trailing 30 days.

11 models · 8 not plotted (no list price)

Developers
0
10
20
30
Effective price per 1M tokens · log scale

Tap a dot to read that model's full profile.

Ranked by developers · last 30 days
  1. 1.Opus 5.5by Anthropic
    28devs
    $0.36 / 1M
  2. 2.GPT 6 1 SOLby OpenAI
    18devs
    no list price
  3. 3.Claude Haiku 5 5by Anthropic
    13devs
    no list price
  4. 4.GPT-6 Astraby OpenAI
    12devs
    $1.31 / 1M
  5. 5.Codex Auto Reviewby OpenAI
    11devs
    $0.47 / 1M
  1. 6.Claude Sonnet 5 5by Anthropic
    10devs
    no list price
  2. 7.Claude Fable 5 2by Anthropic
    8devs
    no list price
  3. 8.Opus 4.8by Anthropic
    7devs
    $0.84 / 1M
  4. 9.GPT-5.6 Solby OpenAI
    6devs
    $0.67 / 1M
  5. 10.Opus 5by Anthropic
    5devs
    $0.86 / 1M

Trailing 30 days. Effective price and cost per turn re-price each model's real token mix (input, output, reasoning and cache tokens) at today's list rates; spend is the cost recorded at sync, as everywhere else on this page. A model needs at least 2 developers and 25Mtokens in the window to be mapped. Per-turn figures leave out Cursor, which reports a whole agent request as a single turn. A model used only through Cursor, like Cursor's own Auto and Composer, has no turns to divide by and isn't mapped.

Effective Price

What a million tokens actually costs once caching is counted: each model's real trailing-30-day token mix, priced at its list rates. The top 12 by spend.

98% of these models' prompt tokens were cache reads.

  • Effective price
  • List price, input → output
  1. Opus 5.598% cache hits
    11× under$0.36
    list $4 in · $20 out
  2. GPT-6 Astra98% cache hits
    7.6× under$1.31
    list $10 in · $50 out
  3. Opus 4.895% cache hits
    5.9× under$0.84
    list $5 in · $25 out
  4. Fable 5.198% cache hits
    14× under$0.71
    list $10 in · $50 out
  5. Codex Auto Review91% cache hits
    5.3× under$0.47
    list $2.50 in · $15 out
  6. Opus 595% cache hits
    5.8× under$0.86
    list $5 in · $25 out
  7. GPT-5.6 Sol94% cache hits
    6.0× under$0.67
    list $4 in · $20 out
  8. GPT-6 Sol98% cache hits
    7.8× under$0.26
    list $2 in · $10 out
  9. GPT-5.6 Terra92% cache hits
    4.9× under$0.41
    list $2 in · $12 out
  10. Auto94% cache hits
    3.6× under$0.35
    list $1.25 in · $6 out
  11. GPT-6 Luna96% cache hits
    6.2× under$0.016
    list $0.10 in · $0.50 out

Effective price = the model's 30-day input, output, reasoning and cache tokens, each billed at its own list rate, divided by the total. “Under” compares it with the list input price. Shows the 12 models with the most spend among those with at least 50M tokens in the window.

Weekly Rhythm

Which days the network codes on: tokens by day of the week over the last 12 whole weeks, split by tool.

Weekends carry 24% of the network's tokens. An even spread would give them 29%. Codex does 61% of weekend tokens, against 40% on weekdays.

Avg tokens per day
0
5B
10B
15B
  1. Monday: 10.8B tokens on an average day, 18% of the week. Claude Code 55%, Codex 43%, Cursor 0.7%, OpenCode 0.1%, Others 1.5%.
  2. Tuesday: 9.5B tokens on an average day, 16% of the week. Claude Code 68%, Codex 30%, Cursor 1.2%, OpenCode 0.3%, Others 0.5%.
  3. Wednesday: 8.4B tokens on an average day, 14% of the week. Claude Code 64%, Codex 32%, Cursor 1.6%, OpenCode 0.4%, Others 1.2%.
  4. Thursday: 7.7B tokens on an average day, 13% of the week. Claude Code 62%, Codex 37%, Cursor 1.2%, OpenCode 0.2%, Others 0.5%.
  5. Friday: 9.3B tokens on an average day, 16% of the week. Codex 58%, Claude Code 40%, Cursor 1.4%, OpenCode 0.3%, Others 0.5%.
  6. Saturday: 5.2B tokens on an average day, 8.6% of the week. Codex 52%, Claude Code 45%, Cursor 1.5%, OpenCode <0.1%, Others 2.0%.
  7. Sunday: 8.9B tokens on an average day, 15% of the week. Codex 66%, Claude Code 32%, Cursor 0.6%, OpenCode <0.1%, Others 0.7%.
Share of tokens, Aug 27 – Nov 18, 2026
Weekend share · by toolNetwork average 24%
  1. 1Codex323B tokens
    32%
    27 devs
  2. 2Antigravity6.6B tokens
    30%
    3 devs
  3. 3Cursor8B tokens
    20%
    5 devs
  1. 4Claude Code378B tokens
    16%
    33 devs
  2. 5OpenCode1.4B tokens
    6.3%
    5 devs

Days are each developer's own calendar days (Cursor's cloud usage is dated in UTC), counted over the 12 whole weeks ending yesterday (Aug 27 – Nov 18, 2026), so every weekday appears exactly 12times and a half-synced today can't dent a column. The chart names, and the weekend list ranks, only what at least 3 developers used in the window. The rest folds into Others, which always stands for at least as many; the tick marks the network average.

Context Length

Crunching individual sessions: how much each model call has to read, prompt size per turn, across sessions started in the last 60 days.

Cursor and Grok Build are left out: they report each request or prompt with all its agent steps rolled into one number, so their “prompt” isn't a single model call.

Top Tools

The coding tools developers run their tokens through: what each one is, how many people use it, and the model it leans on.

  1. 1.Claude Code
    32.1B
    up 16%

    Anthropic's agentic coding CLI

    24 developers · mostly Opus 5.5 (86%)

  2. 2.Codex
    9.7B
    down 56%

    OpenAI's coding agent, CLI and IDE

    13 developers · mostly GPT 6 1 SOL (89%)

  3. 3.Antigravity
    1.2B
    up 5.5%

    Google's agent-first IDE and CLI

    3 developers · mostly Gemini 3 8 Flash N (82%)

  1. 4.OpenCode
    128M
    down 55%

    Open-source terminal coding agent

    1 developer · mostly Deepseek V4 1 Flash Free (100%)

  2. 5.Cursor
    69.6M
    down 72%

    AI code editor forked from VS Code

    1 developer · mostly Grok BOT Automation (71%)

  3. 6.Gemini CLI
    38.5K
    down 96%

    Google's open-source terminal agent

    1 developer · mostly Gemini 3.8 Flash (100%)

No usage this week:

Tokens over the trailing 7 days, across every model a tool ran, against the 7 days before. “New” marks a tool whose first usage falls inside the period; “Back”, one returning after a quiet stretch. Cursor's editor and cloud usage count as one tool, as do Antigravity's app and CLI. Each dot is the tool's colour in Market Share by tool; grey means it folds into Others there.

How these rankings are measured

What the numbers are, where they come from, and what they can't tell you.

Where the numbers come from

Every figure is real usage that developers synced to TokenMaxxer. It comes from the logs Claude Code, Codex, Cursor and the rest already write on their machines, or, for Cursor, from its cloud usage API when they connect it. Tokens count input, output, reasoning, cache reads and cache writes. No prompt or code ever reaches TokenMaxxer. Project names are synced but stay private unless their owner chooses to show them, and nothing here is broken down by project.

Who is counted

Only developers listed on the public leaderboard, and listing is opt-in. It's the same population as the leaderboard and the model, tool and lab pages, so totals reconcile with them, and anyone who leaves the leaderboard leaves every ranking at the same moment.

Time windows

This week and this month are the trailing 7 and 30 days ending today, compared with the 7 or 30 days before them. Weekly charts use the same trailing weeks, so the last column is always this week. Trending ranks week-over-week growth among models with at least 25M tokens this week. A model is new when its first usage (on the network, or inside the tool you picked) falls in the period, and back when it was used before but not in the comparison window.

Why the charts open on six months

Top Models shows the last 13, 26 or 52 weeks and opens on 26; Market Share always covers the trailing 26weeks. Older weeks are lopsided: Claude Code deletes old transcripts from disk while Cursor's cloud sync backfills years, so the far end of a year-long chart mostly shows which tools keep history, not what developers used.

Weekly rhythm

The weekday chart covers 12 whole weeks ending yesterday (Aug 27 – Nov 18), so every weekday is counted the same number of times and a half-synced today can't dent one column; the absolute view shows an average day. Days are each developer's own calendar day. The network's weekend share is Saturday and Sunday tokens over all tokens. An even week would give 2 in 7, about 29%. A tool, model or lab needs at least 3 developers to be named in the chart or listed by weekend share, and Others always stands for at least as many.

Turns and sessions

A turn is one model call (one response from the model), not one prompt you typed; an agent working through a task makes dozens. A session is one conversation in one tool, as the tool records it. Session figures count sessions that started in the last 60 days and went quiet at least 6 hours ago, so one still running is never counted short. Each session is credited to the model behind most of its cost.

Sample floors

A figure built on a handful of sessions describes one person's afternoon, so it is withheld rather than printed thin. A session bucket needs 5 sessions, a model 10, a tool 25 and a context row 500 turns, each from at least 2 developers. The price map lists models with 2+ developers and 25M+ tokens over 30 days; effective price needs 50M tokens.

Prices

Spend is tokens priced at each model's public API list rate, whatever plan the developer actually pays for, as recorded when the usage synced. Effective price and cost per turn instead re-price each model's last 30 days of tokens at today's list rates, so a model priced after its usage arrived doesn't look free. Unpriced sessions are left out of cost figures rather than counted as $0.

Cursor and Grok Build

Cursor reports usage per request and Grok Build per prompt, with every agent step they ran rolled into one number. Their tokens, spend and developers count everywhere, but they sit out of cost per session, session length, context length and the per-turn figures on the price map, where one request would pass for one turn, so a model used only through them isn't mapped. Cursor's cloud sync dates usage by UTC day, which is close enough for weekday totals.

Freshness

Usage data runs through Nov 19, 2026. Rankings refresh every few minutes and session statistics hourly. Devices sync on their own schedule, so the most recent day can still grow as late syncs land.

What these rankings can't tell you

They measure adoption, not quality: a model that re-reads a large context burns more tokens without being better, and a cheap model can lead on volume because it is cheap. They describe developers who chose to share their usage, not the whole market, and on a network this size one heavy user can move a week.