The free tier is closed for most users.Thank you for every chat along the way. Paid models work as before, and your account, API keys, and balance stay as they are.Read why

AI model rankings

Live rankings from real traffic on FreeTheAI. Models are ranked by the tokens they served, with speed, success rate, and share by maker. Browse all models

Usage data through Nov 21, 2026

Requests served
8.3M
Saved by users
$51,716
Active users, 30 days
5,090
Models live
184

Tokens per day by model

The five models with the most tokens in the last 7 days, the rest as Other. Click a model to keep it lit.

  • Claude Sonnet 567.3M
  • GLM-5.339.5M
  • GLM-5.3 Flash10.9M
  • Claude Opus 510.7M
  • MiniMax M33.6M
  • Others13.8M
Tokens served in the last 7 days, by model
DayClaude Sonnet 5GLM-5.3GLM-5.3 FlashClaude Opus 5MiniMax M3OthersTotal
Nov 153.3M6.5M350.1K0418.2K1.1M11.6M
Nov 1610.8M5.6M1.1M0268.8K1.7M19.5M
Nov 179.7M4.7M542.6K0701K1.5M17.1M
Nov 1813.3M5.5M233.6K280.2K176.2K2M21.4M
Nov 1918M6.7M2.2M1.9M621.7K3.4M32.8M
Nov 208.7M9.5M6.2M8.6M1.4M3.4M37.8M
Nov 213.4M917.3K277K067.4K786.8K5.5M

Leaderboard

Models by tokens served in the last 7 days. Open one to compare its offers.

Download JSONCSV

RankModel
  1. 1Claude Sonnet 5Anthropic, 1M context$0.22/M69.6M tokens97.4% success69.6M46.0%up 42056%97.4%13s94 tok/s$0.22/M/ $1.12/M
  2. 2GLM-5.3Zhipu AI, 1M contextFree46.1M tokens99.9% success46.1M30.5%down 99%99.9%24s54 tok/sFree
  3. 3GLM-5.3 FlashZhipu AI, 1M contextFree11.2M tokens100% success11.2M7.4%down 98%100%3.4s69 tok/sFree
  4. 4Claude Opus 5Anthropic, 1M context$0.56/M10.7M tokens93.1% success10.7M7.1%new93.1%58s93 tok/s$0.56/M/ $2.80/M
  5. 5MiniMax M3MiniMax, 1M contextFree3.7M tokens99.8% success3.7M2.5%down 99%99.8%5.8s88 tok/sFree
  6. 6GPT-4o MiniOpenAI, 128K contextFree2.2M tokens99.6% success2.2M1.5%up 1214%99.6%9.3s—Free
  7. 7DeepSeek Flash LatestDeepSeek, 1M context$0.10/M2.1M tokens99.3% success2.1M1.4%new99.3%9.3s191 tok/s$0.10/M/ $1.00/M
  8. 8MiniMax M2.5MiniMax, 128K contextFree846.6K tokens65.4% success846.6K0.6%up 11078%65.4%11s158 tok/sFree
  9. 9Nemotron 3 Ultra 550B A55BNVIDIA, 1M contextFree809K tokens100% success809K0.5%down 98%100%12s51 tok/sFree
  10. 10Claude Sonnet 5.5Anthropic, 1M context$1.00/M742.8K tokens100% success742.8K0.5%down 66%100%13s118 tok/s$1.00/M/ $5.00/M
  11. 11Gemini 4 ArgonGoogle, 1M context$0.22/M578.8K tokens100% success578.8K0.4%new100%46s250 tok/s$0.22/M/ $1.12/M
  12. 12DeepSeek V4.1 FlashDeepSeek, 1M context$0.11/M562.5K tokens32.2% success562.5K0.4%new32.2%6.4s250 tok/s$0.11/M/ $0.57/M
  13. 13Claude Haiku 4.5Anthropic, 200K context$0.92/M435K tokens100% success435K0.3%new100%8.4s199 tok/s$0.92/M/ $4.61/M
  14. 14GPT-5.6 LunaOpenAI, 1.1M context$0.10/M411.1K tokens96.9% success411.1K0.3%new96.9%4.4s250 tok/s$0.10/M/ $1.00/M
  15. 15DeepSeek V4 FlashDeepSeek, 1M context$0.12/M296.1K tokens100% success296.1K0.2%new100%9.8s163 tok/s$0.12/M/ $0.48/M
  16. 16GPT OSS-120BOpenAI, 131.1K contextFree261.9K tokens68.3% success261.9K0.2%new68.3%3.9s190 tok/sFree
  17. 17Gemini 3.8 FlashGoogle, 1M context$0.084/M178.4K tokens100% success178.4K0.1%new100%15s140 tok/s$0.084/M/ $0.42/M
  18. 18Kimi K2.7 CodeMoonshot, 262.1K contextFree134.3K tokens100% success134.3K0.1%down 100%100%7.3s150 tok/sFree
  19. 19Kimi K3Moonshot, 1M context$2.40/M93.6K tokens40.0% success93.6K0.1%new40.0%45s80 tok/s$2.40/M/ $12.00/M
  20. 20GLM-5.3 FlashZhipu AI, 1M context$0.12/M53.9K tokens66.7% success53.9K0.0%up 270%66.7%15s110 tok/s$0.12/M/ $0.40/M

Records

Who leads on use, success, and speed in the last 7 days.

Most tokens served

69.6M

Claude Sonnet 5

Top success rate

100.0%

GLM-5.3 Flash

Fastest median latency

750 ms

Nemotron 3.5 Lightning

Market share by maker

How the top makers compare in the last 7 days: share, success, and speed.

20406080100Token shareRequest shareSuccessSpeedQuick replies
Market share by maker
Token shareRequest shareSuccessSpeedQuick replies
Anthropic54%5.8%97.1%99 tok/s18 s
Zhipu AI38%57%99.8%57 tok/s20 s
MiniMax3.0%6.0%96.0%98 tok/s6.4 s
OpenAI2.0%17%98.3%223 tok/s8.3 s
DeepSeek1.9%3.0%89.2%210 tok/s5.8 s
  1. Anthropic54%
  2. Zhipu AI38%
  3. MiniMax3.0%
  4. OpenAI2.0%
  5. DeepSeek1.9%
  6. Others1.2%

Top clients

By requests, all time, including the original free API.

20406080100RequestsTokens per requestReply lengthCache hitsModels used
Top clients
RequestsTokens per requestReply lengthCache hitsModels used
janitor2.8M24K1.7K51%63
node1.2M53.6K2.6K46%139
openai-sdk513.9K94.3K1.1K84%119
bun362.4K84.7K4.4K51%100
tavo275.7K33.4K2.6K36%87
goonpai9.1K10.9K1.1K67%19
  1. janitor2.8M
  2. node1.2M
  3. openai-sdk513.9K
  4. bun362.4K
  5. tavo275.7K
  6. goonpai9.1K

Daily usage

34.9B tokens and 963.4K requests in the last 30 days.

Each line is scaled to its own 30-day peak · tap Tokens or Requests to see real numbers · drag across the chart to zoom in
Tokens and requests per day over the last 30 days: 34,902,332,550 tokens and 963,430 requests in total
TokensRequests
Oct 231.5B42K
Oct 241.6B43.6K
Oct 251.8B47K
Oct 261.8B47.7K
Oct 271.6B44.3K
Oct 281.5B44K
Oct 291.5B44.1K
Oct 301.8B48.5K
Oct 311.7B49.7K
Nov 12B55K
Nov 22B54.1K
Nov 31.9B51.4K
Nov 42B52.5K
Nov 51.8B49.6K
Nov 62B54K
Nov 72.1B56.2K
Nov 82.2B58.3K
Nov 91.9B52.4K
Nov 101.8B47.8K
Nov 11283.1M10.7K
Nov 128.3M623
Nov 138.6M671
Nov 1410.6M475
Nov 1511.6M765
Nov 1619.5M1.4K
Nov 1717.1M1.2K
Nov 1821.4M1.1K
Nov 1932.8M1.4K
Nov 2037.8M2.5K
Nov 215.5M273

All time: 234.2B input tokens (181.8B of them cached) and 12.1B output tokens. Daily numbers cover FreeTheAI traffic. All-time totals also include the original free API.

People per day by model

Accounts that used each model each day, last 30 days. The legend shows each model's daily average.

  • GLM-5.3764
  • Kimi K2.7 Code182
  • GLM-5.3 Flash151
  • MiniMax M394
  • DeepSeek Chat40
  • Others2.2K
Accounts using each model each day over the last 30 days
DayGLM-5.3Kimi K2.7 CodeGLM-5.3 FlashMiniMax M3DeepSeek ChatOthersTotal
Oct 23966190190141381.8K3.4K
Oct 241K213184137351.9K3.5K
Oct 25996188192118361.8K3.3K
Oct 261K192183115382.1K3.7K
Oct 271K237251166432.4K4.1K
Oct 281.1K235243149442.5K4.3K
Oct 291.1K242265132472.3K4.1K
Oct 301.1K255197133522.4K4.1K
Oct 311.2K277253159603.9K5.8K
Nov 11.2K272244135594K6K
Nov 21.2K285234146443.5K5.4K
Nov 31.2K263220116402.5K4.4K
Nov 41.2K294226144713K5K
Nov 51.2K352269159882.9K5K
Nov 61.2K315224138673.2K5.1K
Nov 71.2K301203131713.2K5.1K
Nov 81.2K313217134753.4K5.4K
Nov 91.2K334231138773.9K5.9K
Nov 101.2K3682631651063.2K5.3K
Nov 1179224212275542.5K3.8K
Nov 12108212417131.2K1.4K
Nov 137215159101.5K1.6K
Nov 1461111355678773
Nov 155013171151.3K1.4K
Nov 1641915961.6K1.7K
Nov 173685861.6K1.7K
Nov 1834111164448514
Nov 193357559791K
Nov 202681054885938
Nov 2153321573587

Retention and cache hits

Where people use FreeTheAI

Accounts by the country of their latest request in the last 30 days.

Accounts
9,532
Countries
145

Top countries

  1. United States738
  2. Indonesia302
  3. India270
  4. Brazil266
  5. Philippines249
  6. Netherlands231
  7. Germany224
  8. China185
  9. Russia117
  10. Vietnam105

2,403 more users in 135 other countries.

Countries with fewer than 3 users are counted together. Updated every 30 minutes.

How we count

Only our own traffic
Every request through the FreeTheAI gateway is counted once, in 5-minute, hourly, and daily buckets on UTC. No vendor numbers, benchmarks, or votes are used, and only models in the public catalog are named.
Tokens and ranking
Tokens are input, cached ones included, plus output, as the provider reported them. Models rank by tokens served, and a tie goes to more requests. Share is a model's part of all tokens in the window.
Growth
Growth compares each model with the window just before it, using finished hours for 24 hours and finished days otherwise, so a day in progress never looks like a drop. No earlier traffic means no figure.
Success and speed
Success counts what the provider answered or failed; a caller's own bad request never counts. Speed is output tokens per second on streamed replies, shown as the median. Records need 20 requests in the window.
People and retention
A person is one FreeTheAI account, counted once per model per day. Retention compares the week before last with last week, for models that had at least 20 people in the first of those weeks.
Cache hits and prices
Cache hits are input tokens the provider read from its prompt cache over the last 7 days. Prices are what FreeTheAI charges per million tokens, beside the model maker's own published price.

Frequently asked questions

How is the leaderboard ranked?

By tokens FreeTheAI served for each model in the selected window. It is a measure of real use, not a quality score.

How fresh is it?

Requests are counted within about a minute, and the page refreshes its numbers every minute.

Why is a model missing?

A model appears once it is listed in the catalog and has served at least one request in the window.

Why does a free model rank so high?

Free models get a lot of traffic. Use the price filter to compare free and paid models separately; ranks stay the same so you can see where each sits overall.

Can I use this data?

Yes. Download it as JSON or CSV from the ranking section. Please credit FreeTheAI.

Includes 176.2B tokens and 6.4M requests from the original free API, as of Sep 29, 2026, 1:18 AM UTC.

Updated Nov 21, 2026, 2:43 AM UTC. Numbers refresh every minute.