OverviewAll LeaderboardsMethodology
New Chat

Vision Arena

What is the best model for vision tasks?

Nov 8, 2026
813,798 votes
122 models

Vision Arena

Frontier
AnthropicClaude Fable 5 (High)1,327 rating. 11 points above rank 2.
Daily Driver
Gemini 3.8 Flash (High)Near-frontier performance at 1/10th the price.
Efficient
Gemma 4 31BSolid performance at 1/100th the price.
1,2001,300
  1. 1,327
    Anthropic
    Claude Fable 5High
  2. 1,316
    Qwen3.8-Max
  3. 1,316
    Anthropic
    Claude Opus 4.7
  4. 1,315
    Gemini 3.8 FlashHigh
  5. 1,313
    Anthropic
    Claude Opus 4.6High
  6. 1,312
    Anthropic
    Claude Opus 5.5High
  7. 1,312
    Anthropic
    Claude Opus 4.6
  8. 1,311
    Anthropic
    Claude Opus 4.7High
  9. 1,309
    Meta
    Muse Spark 1.3Max
  10. 1,308
    Gemini 3.7 FlashHigh
  11. 1,306
    Meta
    Muse Spark 1.2xHigh
  12. 1,305
    Anthropic
    Claude Fable 5.1Max
  13. 1,303
    Meta
    Muse Spark
  14. 1,302
    Gemini 3 Pro
  15. 1,301
    Meta
    Muse Spark 1.3xHigh
Model RankingScoreCost per task (P50)Blended price $/M (3:1)
  1. 1AnthropicClaude Fable 5 (High)1,327$0.21$40
  2. 2Qwen3.8-Max1,316$0.04$5
  3. 3AnthropicClaude Opus 4.71,316$0.09$20
  4. 4Gemini 3.8 Flash (High)1,315$0.02$3
  5. 5AnthropicClaude Opus 4.6 (High)1,313$0.10$20
  6. 6AnthropicClaude Opus 5.5 (High)1,312$0.10$16
  7. 7AnthropicClaude Opus 4.61,312$0.09$20
  8. 8AnthropicClaude Opus 4.7 (High)1,311$0.10$20
  9. 9MetaMuse Spark 1.3 (Max)1,309$0.03N/A
  10. 10Gemini 3.7 Flash (High)1,308$0.02$3
  11. 11MetaMuse Spark 1.2 (xHigh)1,306$0.03$3.50
  12. 12AnthropicClaude Fable 5.1 (Max)1,305$0.44$40
  13. 13MetaMuse Spark1,303N/AN/A
  14. 14Gemini 3 Pro1,302N/A$9.50
  15. 15MetaMuse Spark 1.3 (xHigh)1,301$0.03N/A

Pareto frontier

Best score for the cost
1,2001,2501,300
Anthropic
$0.001$0.01$0.10

Median cost per task

Frontier
AnthropicClaude Fable 5 (High)1,327 rating. 11 points above rank 2.
Daily Driver
Gemini 3.8 Flash (High)Near-frontier performance at 1/10th the price.
Efficient
Gemma 4 31BSolid performance at 1/100th the price.
On the frontierScore$/task
  1. AnthropicClaude Fable 5 (High)1,327$0.21
  2. Qwen3.8-Max1,316$0.04
  3. Gemini 3.8 Flash (High)1,315$0.02
  4. Grok 4.51,290$0.01
  5. GLM-5.3-Flash1,287$0.01
  6. Gemma 4 31B1,277$0.01
52 models not plotted

These models are on the leaderboard but do not appear on the Pareto chart because they have insufficient cost data. The numbers beside their names show their leaderboard ranks.

  • 13Muse Spark
  • 14Gemini 3 Pro
  • 36Gemini 3 Flash
  • 48Dola Seed 2.0 Pro
  • 56GPT-5.1 (High)
  • 60GPT-5.2 (High)
  • 62Gemini 2.5 Pro
  • 64GPT-5.1
  • 69ChatGPT-4o Latest (20250326)
  • 70GPT-5 Chat
  • 73Kimi K2.5 (Instant)
  • 75Gemini 2.5 Flash Preview (09-2025)
  • 76GPT-5.2
  • 77GPT-5 (High)
  • 79ERNIE-5.0-Preview (1220)
  • 80Qwen3-VL-235B-A22B-Instruct
  • 82GPT-4.1 (2025-04-14)
  • 83o3 (2025-04-16)
  • 84NVIDIA-Nemotron-3.5-Super-VL-120B-A12B (NVFP4)
  • 85Gemini 2.5 Flash
  • 87Claude Opus 4 (Thinking 16K, 20250514)
  • 88Claude Sonnet 4 (Thinking 32K, 20250514)
  • 91Qwen-VL-Max (2025-08-13)
  • 93o4-mini (2025-04-16)
  • 94Claude 3.7 Sonnet (Thinking 32K, 20250219)
  • 95GPT-4.1 mini (2025-04-14)
  • 96GPT-5 mini (High)
  • 97Qwen3-VL-235B-A22B-Thinking
  • 98Claude Opus 4 (20250514)
  • 100Claude Sonnet 4 (20250514)
  • 101Gemini 2.5 Flash-Lite Preview (Thinking, 06-17)
  • 102Claude 3.7 Sonnet (20250219)
  • 103Gemini 2.5 Flash-Lite Preview (Non-Thinking, 09-2025)
  • 104Claude 3.5 Sonnet (20241022)
  • 105Grok 4 (0709)
  • 106Mistral Medium 3.1 (2508)
  • 107GLM-4.6V
  • 108Gemini 2.0 Flash (001)
  • 109Gemma 3 27B IT
  • 110Mistral Medium 3 (2505)
  • 111GPT-5 nano (High)
  • 112GLM-4.5V
  • 113Hunyuan Vision 1.5 Thinking
  • 114Llama 4 Maverick 17B 128E Instruct
  • 115Step-1o Turbo (202506)
  • 116Step-3
  • 117Hunyuan Large Vision
  • 118Mistral Small 3.2 (2506)
  • 119Mistral Small 3.1 24B Instruct (2503)
  • 120Claude 3.5 Haiku (20241022)
  • 121Llama 4 Scout 17B 16E Instruct
  • 122Molmo 2 8B

Default Leaderboard Plots

1 model with no battles against the other models is hidden.

Previous
Search

Which model is the best at online search?

Next
Document

Which model is the best at analyzing documents?

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