Skip to content
OSMEXO logoOSMEXO

The AI model market, today

Demand, momentum, benchmarks and what people are saying, refreshed from public sources.

Demand

Daily tokens by model. Hover a day to see every model.

Last 7 days292.5T tokens
vs previous 7 days▲ 52%
21 Nov vs day before▼ 33%
tokens per day
011.3T22.6T33.9T45.1T15 Nov17 Nov19 Nov21 Nov
Daily tokens, 7 days
DayHy4 flashDeepSeek V4.1 Flash 1017DeepSeek V4.1 FlashGLM 5.3 FlashGPT-6 LunaHy4Claude Opus 5.5MiniMax M3.1 FlashTop 48 models
15 November 20260 tokens7.2T tokens1.9T tokens1.2T tokens918.5B tokens1.9T tokens690.6B tokens535.5B tokens23.4T tokens
16 November 20261.9T tokens9.1T tokens2.2T tokens1.4T tokens1.2T tokens1.1T tokens1.3T tokens532.8B tokens29.7T tokens
17 November 202620.9T tokens9.4T tokens2.7T tokens1.8T tokens1.4T tokens1.1T tokens1.3T tokens564.2B tokens51.0T tokens
18 November 202623.8T tokens9.8T tokens2.5T tokens1.9T tokens1.5T tokens1.3T tokens1.1T tokens560.1B tokens53.8T tokens
19 November 202622.6T tokens9.6T tokens2.2T tokens1.5T tokens1.6T tokens1.4T tokens1.1T tokens671.5B tokens51.4T tokens
20 November 202623.4T tokens9.1T tokens2.1T tokens1.6T tokens1.1T tokens821.5B tokens976.1B tokens705.7B tokens49.7T tokens
21 November 202610.3T tokens9.0T tokens1.7T tokens2.2T tokens761.6B tokens48.9B tokens557.9B tokens630.8B tokens33.5T tokens

The 7-day view covers the top 48 models, day by day. The 8 biggest models are drawn; smaller ones are left out of the stacks. Headline numbers and the tooltip total count everything.

DeepSeek V4 Flash
DeepSeek · 4 models in top 100
83.2T
tokens / week
▲ 9.2% vs last week
GLM 5 Flash
Z.ai · 1 model in top 100
11.6T
tokens / week
▼ 2.0% vs last week
GPT-6 Luna
OpenAI · 1 model in top 100
8.6T
tokens / week
▲ 34% vs last week
Claude Opus 5
Anthropic · 2 models in top 100
7.2T
tokens / week
▲ 47% vs last week
Hy4 flash
Tencent · 1 model in top 100
6.5T
tokens / week
New this week
JEV 1
Typesafe · 2 models in top 100
5.2T
tokens / week
▲ 6.9% vs last week
GLM 5
Z.ai · 2 models in top 100
4.4T
tokens / week
▼ 7.0% vs last week
MiniMax M3 Flash
MiniMax · 1 model in top 100
4.2T
tokens / week
▲ 442% vs last week
Kimi K3 Pro
Moonshot AI · 1 model in top 100
3.9T
tokens / week
▼ 1.1% vs last week
Step 5
StepFun · 1 model in top 100
3.8T
tokens / week
▲ 211× vs last week

Families of the top 100 models; daily tokens over 14 days.

Why each model is here, with the number behind it

#ModelWhy trendingTokens / week14 daysBest forContext · list priceLiveBench
1
DeepSeek V4.1 Flash
DeepSeek · Open weights
Demand #299 mentionsNew this quarter
15.3T
▼ 2.3%
Vision tasks
1M
$0.05 / $0.20
—
2
GLM 5.3 Flash
Z.ai · Open weights
Demand #341 mentionsNew this quarter
11.6T
▼ 2.0%
Design · #23
1M
$0.04 / $0.08
60.7
3
—
70.9B
▲ 45%
—Vision tasks
1M
$1.50 / $9.00
—
4
Step 5 Preview
StepFun · Open weights
+20976% usageDemand #94 mentions
3.8T
▲ 211×
Reasoning · #26
1M
$1.00 / $2.70
54.9
5
Demand #93
91.9B
▼ 15%
—Vision tasks
1M
$2.00 / $12.00
—
6
Hy4 flash
Tencent · Open weights
Demand #6266 mentionsNew this month
6.5T
new
Vision tasks
1M
$0.04 / $0.14
—
7
Inkling 2 Mini
Thinking Machines · Open weights
Trending with developersNew this month116 mentions
41.6B
new
—Vision tasks
1M
$0.07 / $0.30
—
8
Kimi K3
Moonshot AI · Open weights
165 mentionsDemand #32
997.0B
▼ 14%
Reasoning · #11
1M
$0.19 / $16.00
65.0
9
Demand #57
383.6B
▼ 19%
—Vision tasks
1M
$0.03 / $0.13
—
10
MiniMax M3.1 Flash
MiniMax · Open weights
Demand #7+442% usage11 mentions
4.2T
▲ 442%
Browser · #13
1M
$0.15 / $0.60
63.7
11
488 mentionsNew this month+303% usage
450.2B
▲ 303%
—Data Analysis · #6
1M
$1.50 / $7.50
77.0
12
Kimi K3.1 Pro
Moonshot AI · Open weights
Demand #8126 mentionsNew this month
3.9T
▼ 1.1%
Reasoning · #10
1M
$0.25 / $8.80
68.7
13
GLM 5.3
Z.ai · Open weights
Demand #1370 mentions
3.4T
▼ 7.2%
—
1M
$0.07 / $1.40
—
14
268 mentionsNew this monthDemand #46
578.5B
new
Research · #7
1M
$2.00 / $10.00
80.3
15
MiMo-V2.6-Flash
Xiaomi · Open weights
+375% usageDemand #1010 mentions
3.8T
▲ 375%
Vision tasks
1M
$0.14 / $0.28
—
16
120 mentionsNew this month
———
1M
—
—
17
514 mentionsNew this monthDemand #55
391.9B
new
—Browser · #1
1M
$3.00 / $15.00
87.4
18
Gemma 3 27B
Google · Open weights
—
18.0B
▼ 43%
—Vision tasks
128K
$0.08 / $0.45
—
19
281 mentions+333% usageNew this month
2.2T
▲ 333%
Data Analysis · #1
1M
$2.00 / $10.00
83.3
20
Qwen4 Max
Alibaba
367 mentionsNew this monthDemand #27
1.3T
new
Reasoning · #4
1M
$1.43 / $4.28
73.1

Under the context window: the list price per 1M input / output tokens. Best for = the model's strongest LiveBench task type, with its rank.

Latest news & research

Official posts, papers and release notes

By the numbers

Facts computed from this week's data, each with its source

35.6×

more tokens a week than a year ago: 211.3T now vs 5.9T.

75%

of tokens go to open-weight models, up from 28% a year ago.

Open 75% · Closed 25%

39%

of this week's tokens went to DeepSeek, the biggest vendor.

DeepSeek 39%, Tencent 10%, OpenAI 9%, Z.ai 8%, Anthropic 7%, NVIDIA 4% of this week's tokens.

16.4×

week-over-week growth for Pplx Decider Flash V1.3 (free) (Perplexity), the fastest climber.

Pplx Decider Flash V1.3 (free)

14.2T

tokens a week used by Hermes Agent, the biggest app.

Hermes Agent14.2T
Claude Code11.2T
Kilo Code8.7T
pi5.5T

87.4

overall LiveBench score for Gemini 4 Neon, the current leader.

Gemini 4 Neon87.4
Claude Fable 5.2 (xhigh)84.9
GPT-6.2 Astra84.9

$0.52

per successful task for GPT-6.2 Sol, the cheapest of LiveBench's top 10.

GPT-6.2 Sol$0.52
GPT-6.1 Terra$0.67
GPT-6.1 Sol$0.82

Benchmarks

LiveBench release 6 Nov 2026. Pick a task type to rank by it.

  1. 1st · Overall
    Gemini 4 Neon
    Google · xhigh effort
    87.4/ 100
    Overall 87.4
    Cost per successful task $2.65
    • best in Overall
    • best in Browser
    • best in Research
  2. 2nd · Overall
    Claude Fable 5.2 (xhigh)
    Anthropic · xhigh effort
    84.9/ 100
    Overall 84.9
    Cost per successful task $9.85
    • best in Long Agentic
    • best in Research
    • best in Design
  3. 3rd · Overall
    GPT-6.2 Astra
    OpenAI · max effort
    84.9/ 100
    Overall 84.9
    Cost per successful task $1.94
    • best in Coding
LiveBench rows 4 to 18, sorted by Overall. Scores out of 100; ★ marks the best in each column.
#ModelOverallLong AgenticReasoningCodingBrowserData AnalysisResearchDesignCost per successful task
4
GPT-6 Astra
OpenAI · max effort
84.3
Long Agentic 74.7Reasoning 92.7Coding 87.5Browser 91.3Data Analysis 83.0Research 90.6Design 80.0
$5.33
5
Claude Fable 5.1
Anthropic · max effort
84.0
Long Agentic 74.2Reasoning 91.7Coding 89.7Browser 89.8Data Analysis 80.3Research 88.5Design 81.6
$8.36
6
Claude Opus 5.5
Anthropic · max effort
83.4
Long Agentic 72.8Reasoning 92.2Coding 87.1Browser 88.4Data Analysis 80.3Research 88.9Design 83.3
$6.72
7
Gemini 4 Carbon
Google · high effort
83.3
Long Agentic 71.8Reasoning 90.6Coding 91.0Browser 80.0Data Analysis 86.5, best in columnResearch 92.5Design 80.2
$2.03
8
GPT-6.1 Sol
OpenAI · max effort
81.2
Long Agentic 64.7Reasoning 92.6Coding 82.1Browser 93.3Data Analysis 82.7Research 90.0Design 79.6
$0.82
9
GPT-6.2 Sol
OpenAI · max effort
80.3
Long Agentic 63.6Reasoning 90.6Coding 83.7Browser 88.2Data Analysis 81.1Research 90.1Design 80.0
$0.52
10
GPT-6.1 Terra
OpenAI · max effort
78.6
Long Agentic 55.3Reasoning 88.7Coding 89.2Browser 93.1Data Analysis 82.3Research 79.1Design 79.7
$0.67
11
Gemini 4 Argon
Google · high effort
77.7
Long Agentic 58.5Reasoning 95.1, best in columnCoding 85.7Browser 82.8Data Analysis 85.2Research 86.8Design 69.2
$1.82
12
Gemini 4 Hydrogen
Google · xhigh effort
77.0
Long Agentic 61.8Reasoning 88.6Coding 78.5Browser 77.1Data Analysis 83.2Research 86.8Design 77.3
$1.92
13
Claude Sonnet 5.5
Anthropic · max effort
76.8
Long Agentic 62.5Reasoning 91.6Coding 77.2Browser 89.3Data Analysis 59.5Research 88.0Design 81.3
$5.82
14
Qwen4-Max
Alibaba · xhigh effort
73.1
Long Agentic 46.6Reasoning 92.2Coding 69.8Browser 80.7Data Analysis 83.4Research 84.7Design 79.1
$0.69
15
Muse Wonder 1.0
Meta · max effort
70.7
Long Agentic 46.2Reasoning 90.6Coding 74.9Browser 60.0Data Analysis 81.6Research 93.0Design 70.9
$2.46
16
DeepSeek V4.1 Flash 1017
DeepSeek · max effort
70.3
Long Agentic 38.8Reasoning 89.1Coding 80.4Browser 67.2Data Analysis 79.5Research 78.6Design 79.8
$0.19
17
Claude Haiku 5.5
Anthropic · max effort
68.9
Long Agentic 47.9Reasoning 82.4Coding 77.6Browser 69.5Data Analysis 53.1Research 84.3Design 79.3
$1.26
18
Kimi K3.1 Pro
Moonshot AI · max effort
68.7
Long Agentic 38.7Reasoning 91.1Coding 67.8Browser 75.9Data Analysis 79.1Research 82.1Design 73.7
$1.07

Cell colour deepens with the score; the number is always shown. ★ marks the best in each column. Hover a cell, or focus it and use the arrow keys, for its sub-areas.

Model universe

Vendors, families and models, sized by weekly tokens. Drag nodes, Ctrl/⌘ + scroll or pinch to zoom, click to inspect.

Top 50 models this week, by vendor

  • DeepSeek: 84.6T tokens this week
    • DeepSeek V4.1 Flash 1017, #1, 63.4T tokens
    • DeepSeek V4.1 Flash, #2, 15.3T tokens
    • DeepSeek V4 Flash 0731, #19, 2.6T tokens
    • DeepSeek V4 Flash, #23, 2.0T tokens
    • DeepSeek V4 Pro 0813, #39, 734.5B tokens
    • DeepSeek V4 Pro, #44, 594.7B tokens
  • Z.ai: 16.0T tokens this week
    • GLM 5.3 Flash, #3, 11.6T tokens
    • GLM 5.3, #13, 3.4T tokens
    • GLM 5.2, #35, 906.8B tokens
  • OpenAI: 14.9T tokens this week
    • GPT-6 Luna, #4, 8.6T tokens
    • GPT-6.1 Sol, #20, 2.6T tokens
    • GPT-5.6 Luna, #24, 1.7T tokens
    • GPT-6.1 Terra, #34, 922.9B tokens
    • GPT-6 Luna Pro, #45, 593.9B tokens
    • GPT-6.2 Sol, #46, 578.5B tokens
  • Anthropic: 13.5T tokens this week
    • Claude Opus 5.5, #5, 7.1T tokens
    • Claude Haiku 5.5, #17, 2.9T tokens
    • Claude Sonnet 5.5, #18, 2.8T tokens
    • Claude Fable 5.2, #40, 713.8B tokens
  • NVIDIA: 8.4T tokens this week
    • Nemotron 3 Ultra, #11, 3.8T tokens
    • Nemotron 3.5 Super VL, #14, 3.4T tokens
    • Nemotron 3.5 Lightning, #31, 1.2T tokens
  • Google: 6.9T tokens this week
    • Gemini 4 Carbon, #22, 2.2T tokens
    • Gemini 3.8 Flash, #25, 1.7T tokens
    • Gemini 4 Argon, #41, 695.2B tokens
    • Gemini 3 Flash Preview, #42, 651.8B tokens
    • Gemma 4 26B A4B IT, #43, 644.8B tokens
    • Gemini 3.1 Flash Lite, #48, 529.1B tokens
    • Gemma 4 31B IT, #49, 518.0B tokens
  • Tencent: 6.5T tokens this week
    • Hy4 flash, #6, 6.5T tokens
  • Typesafe: 5.2T tokens this week
    • JEV 1.13 (TypeSafe), #16, 2.9T tokens
    • JEV 1.14 (TypeSafe), #21, 2.3T tokens
  • Moonshot AI: 4.9T tokens this week
    • Kimi K3.1 Pro, #8, 3.9T tokens
    • Kimi K3, #32, 997.0B tokens
  • MiniMax: 4.8T tokens this week
    • MiniMax M3.1 Flash, #7, 4.2T tokens
    • MiniMax M3, #47, 555.6B tokens
  • Xiaomi: 4.7T tokens this week
    • MiMo-V2.6-Flash, #10, 3.8T tokens
    • MiMo-V2.6-Pro, #37, 864.5B tokens
  • Meta: 4.1T tokens this week
    • Muse Spark 1.3 Contributor, #15, 3.1T tokens
    • Muse Ember, #33, 942.5B tokens
  • StepFun: 3.8T tokens this week
    • Step 5 Preview, #9, 3.8T tokens
  • Baidu: 3.5T tokens this week
    • ERNIE-5.5-Flash-Preview, #12, 3.5T tokens
  • Alibaba: 3.3T tokens this week
    • Qwen4 Max, #27, 1.3T tokens
    • Qwen3.8 Flash, #29, 1.2T tokens
    • Qwen3.8 27B, #38, 800.5B tokens
  • Thinking Machines: 1.6T tokens this week
    • Inkling 2 Small, #26, 1.6T tokens
  • Upstage: 1.2T tokens this week
    • Solar Mini 4, #28, 1.2T tokens
  • Particle AI: 1.2T tokens this week
    • Particle-1, #30, 1.2T tokens
  • Dots Studio: 876.3B tokens this week
    • Dots3-Note Preview, #36, 876.3B tokens
  • Perplexity: 479.0B tokens this week
    • Pplx Decider V1.2, #50, 479.0B tokens

Node size: weekly tokens. A family node appears when a family has two or more models in view.

Trends

How the frontier moved, and what each point of intelligence costs

Benchmark data: Artificial Analysis

The best Intelligence Index score each lab had released by each date

Intelligence Index
0204060Jul2024Jul2025Jul2026Jul
  • Google60.0
  • Anthropic59.1
  • OpenAI56.9
  • Moonshot AI55.1
  • Meta52.4
  • xAI46.4
  • Xiaomi46.3
  • DeepSeek45.1

Media leaderboards

Top models by Elo from blind side-by-side votes

Benchmark data: Artificial Analysis
Image leaderboard
#ModelElo95% CIReleased
11,223±10Sep 2026
21,207±10Sep 2026
3
GPT Image 2 (high)OpenAI
1,189±11Apr 2026
41,176±12Oct 2026
5
Grok Imagine Image 2.0SpaceXAI
1,168±12Aug 2026
6
MAI-Image-2.6Microsoft AI
1,167±12Aug 2026
7
Nano Banana 2 (Gemini 3.1 Flash Image)Google
1,140±10Feb 2026
8
Muse ImageMeta
1,127±11Jul 2026
9
GPT Image 1.5 (high)OpenAI
1,123±11Dec 2025
10
MAI-Image-2.6-FlashMicrosoft AI
1,120±12Sep 2026

What people are watching and discussing

The last two weeks, ranked by views and engagement

Latest AI research

New papers from arXiv and the models they discuss

Demand, benchmarks, news, research and social posts are refreshed from public sources. List prices are public prices per 1M tokens.