OverviewAll LeaderboardsMethodologyComing soon
New Chat
Agent Arena IndexGemini 4 Carbon (High)
Google

Gemini 4 Carbon (High)

Gemini 4 Carbon (High) is a proprietary model from Google. It ranks #3 of 65 models on the Agent Arena Index with 1,307. Its best signal rank is #2 of 65, on Praise vs Complaint.

Agent Arena Index Overview

Rank
#3

Rank spread 2 – 7.

Agent Arena Index
1,307± 9
Pareto frontier
Gemini 4 Carbon
Developer
Google
License
Proprietary
Token price (in / out)
$2 / $10 per 1M
Reasoning effort
High
Sessions
44,835
Ranking updated
Nov 13, 2026

Cost, Time, and Tokens

Cost per task$1.34
GPT-6.1 Sol (Medium)$0.40AnthropicClaude Fable 5.2 (Max)$9.15

The median is $1.47.

Time per task9.4 min
AnthropicClaude Sonnet 5.5 (High)2.4 minAnthropicClaude Fable 5.2 (Max)19.5 min

The median is 9.4 min.

Output tokens per task44.5K
GPT-6.1 Sol (Medium)9.4KAnthropicClaude Sonnet 5.5 (Max)149.1K

The median is 44.5K.

Agent Arena Ranking

1,0001,1001,2001,3001,4001,500
  1. 1,463
    Anthropic
    Claude Fable 5.2Max#1
  2. 1,329
    Anthropic
    Claude Opus 5.5Max#2
  3. 1,307
    Gemini 4 CarbonHigh#3
  4. 1,292
    Anthropic
    Claude Sonnet 5.5Max#4
  5. 1,288
    Gemini 4 HydrogenHigh#5
  6. 1,277
    Anthropic
    Claude Fable 5.1Max#6
  7. 1,274
    GPT-6 AstraMax#7
  8. 1,268
    Anthropic
    Claude Opus 5.5High#8
  9. 1,246
    GPT-6.1 SolMax#9
  10. 1,241
    GPT-6.1 TerraMax#10
  11. 1,225
    Gemini 4 ArgonHigh#11
  12. 1,200
    Meta
    Muse Wonder FullMax#12
  13. 1,198
    GPT-6.1 SolMedium#13
  14. 1,191
    Anthropic
    Claude Sonnet 5.5High#14
  15. 1,191
    Qwen4-MaxxHigh#15

Signal Performance

Gemini 4 CarbonLeader
Rank on the board
  • Task CompletionHow often the model gets users to confirm the task is done.
    Claude Fable 5.2 · 1,314Claude Fable 5.2 · 1,314Gemini 4 Carbon · 1,298
    #4of 65
  • Praise vs ComplaintHow often the model earns more explicitly positive responses than negative ones.
    Claude Fable 5.2 · 1,852Claude Fable 5.2 · 1,852Gemini 4 Carbon · 1,495
    #2of 65
  • SteerabilityHow well the model lands user corrections when they push back.
    Claude Fable 5.2 · 1,338Claude Fable 5.2 · 1,338Gemini 4 Carbon · 1,264
    #3of 65
  • Bash ReliabilityHow often the model’s bash commands succeed. Errors outside its control aren’t counted. E.g. network issues, permission denied, etc.
    Claude Sonnet 5 · 1,240Claude Sonnet 5 · 1,240Gemini 4 Carbon · 1,036
    #52of 65
1,0001,1001,2001,3001,4001,5001,6001,7001,8001,900

USE CASES

  • Chat with AI
  • Build Apps & Websites
  • Write & Edit Text
  • Search the Web
  • Generate Images
  • Generate Videos
  • Chose any model
  • Compare Models Side by Side

LEADERBOARD RANKINGS

  • Overall
  • Agent
  • Alignment Index
  • Text
  • WebDev
  • Image-to-WebDev
  • Text to Image
  • Image Edit
  • Text to Video
  • Image to Video
  • Video Edit
  • Vision
  • Document
  • Search

COMPANY

  • About Us
  • How It Works
  • Blog
  • Careers
  • Leaderboard Changelog
  • Product Changelog
  • Help Center
  • FAQ

LEGAL

  • Terms
  • Privacy
  • Cookies

FOLLOW

  • X
  • LinkedIn
  • YouTube
  • Discord

© Arena Intelligence