OverviewAll LeaderboardsMethodologyComing soon
New Chat
Work IndexGPT-5.6 Terra (xHigh)
OpenAI

GPT-5.6 Terra (xHigh)

GPT-5.6 Terra (xHigh) is a proprietary model from OpenAI. It ranks #41 of 65 models on the Work Index with 1,068. Its best signal rank is #33 of 65, on Praise vs Complaint.

Work Index Overview

Rank
#41

Rank spread 39 – 46.

Work Index
1,068± 13
Pareto frontier

128 points under the frontier

GPT-5.6 Terra
Developer
OpenAI
License
Proprietary
Token price (in / out)
$2 / $12 per 1M
Reasoning effort
xHigh
Sessions
18,769
Ranking updated
Nov 13, 2026

Cost, Time, and Tokens

Cost per task$0.80
GPT-6 Luna (Low)$0.01AnthropicClaude Fable 5.1 (Max)$12.25

The median is $0.70.

Time per task3.6 min
MetaMuse Ember (Max)35 sAnthropicClaude Fable 5.2 (Max)25.1 min

The median is 8 min.

Output tokens per task23.2K
GPT-6 Luna (Low)5KAnthropicClaude Sonnet 5.5 (Max)174.1K

The median is 43.4K.

Work Ranking

9001,0001,100
  1. 1,096
    GLM-5.3Max#34
  2. 1,095
    GLM-5.2Max#35
  3. 1,095
    Gemini 3.8 FlashHigh#36
  4. 1,094
    Tencent
    Hy4 preview#37
  5. 1,094
    Kimi K3Max#38
  6. 1,083
    Qwen3.8-Max#39
  7. 1,075
    Grok 4.6xHigh#40
  8. 1,068
    GPT-5.6 TerraxHigh#41
  9. 1,065
    GPT-6 LunaMedium#42
  10. 1,061
    Stepfun
    Step 5 Preview#43
  11. 1,055
    MiniMax-M3.1-Flash#44
  12. 1,047
    GLM-5.3-Flash#45
  13. 1,037
    Qwen3.8-Flash-Next#46
  14. 1,035
    Gemini 3.7 FlashHigh#47
  15. 1,024
    Qwen3.8-27B#48

Signal Performance

GPT-5.6 TerraLeader
Rank on the board
  • Task CompletionHow often the model gets users to confirm the task is done.
    Claude Sonnet 5.5 (Max) · 1,307Claude Sonnet 5.5 (Max) · 1,307GPT-5.6 Terra · 1,042
    #50of 65
  • Praise vs ComplaintHow often the model earns more explicitly positive responses than negative ones.
    Claude Fable 5.2 · 1,861Claude Fable 5.2 · 1,861GPT-5.6 Terra · 1,079
    #33of 65
  • SteerabilityHow well the model lands user corrections when they push back.
    Claude Fable 5.2 · 1,376Claude Fable 5.2 · 1,376GPT-5.6 Terra · 1,073
    #35of 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 · 1,234Claude Sonnet 5 · 1,234GPT-5.6 Terra · 1,085
    #41of 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