Best Multimodal AI Models

Multimodal models take more than text as input. We rank them by overall benchmark intelligence and note context window and pricing for each.

16 models ranked · updated continuously

How we rank: Ranked by benchmark intelligence across text, vision, and audio inputs.

  1. 01

    Gemini 3.5 UltraGoogle DeepMind

    Google's top multimodal model with native video understanding.

    IQ
    1463
    Rating
    4.90 / 5
    Context
    2M tokens
    Price
    $10/M tokens
    License
    proprietary
  2. 02

    Gemini 3 UltraGoogle DeepMind

    4M context, agentic, video-native.

    IQ
    1452
    Rating
    4.90 / 5
    Context
    4M
    Price
    $6/M tokens
    License
    Closed
  3. 03

    Claude Opus 4.5 VisionAnthropic

    Opus with native image + doc understanding.

    IQ
    1430
    Rating
    4.85 / 5
    Context
    500K
    Price
    $15/M tokens
    License
    Closed
  4. 04

    Gemini 3 ProGoogle DeepMind

    2M tokens of context , the long-memory champion.

    IQ
    1402
    Rating
    4.70 / 5
    Context
    2M
    Price
    $2.5/M tokens
    License
    Closed
  5. 05

    GPT-5 VisionOpenAI

    GPT-5 with native vision + voice.

    IQ
    1395
    Rating
    4.80 / 5
    Context
    400K
    Price
    $6/M tokens
    License
    Closed
  6. 06

    Reka Core 2Reka AI

    Native multimodal frontier

    IQ
    1324
    Rating
    4.40 / 5
    Context
    256K
    Price
    $5/M tokens
    License
    proprietary
  7. 07

    Pixtral LargeMistral AI

    Mistral vision flagship

    IQ
    1318
    Rating
    4.40 / 5
    Context
    128K
    Price
    $2/M tokens
    License
    open-weights
  8. 08

    Qwen 3 VLAlibaba

    Open multimodal Qwen

    IQ
    1318
    Rating
    4.40 / 5
    Context
    256K
    Price
    $0.5/M tokens
    License
    open-weights
  9. 09

    Amazon Nova Pro 2Amazon

    AWS native multimodal

    IQ
    1316
    Rating
    4.30 / 5
    Context
    300K
    Price
    $1.6/M tokens
    License
    proprietary
  10. 10

    Qwen 3 VL MaxAlibaba

    Qwen's vision-language flagship.

    IQ
    1310
    Rating
    4.50 / 5
    Context
    1M
    Price
    $1/M tokens
    License
    Open Weights
  11. 11

    Reka Core 3Reka AI

    Open-weights multimodal frontier.

    IQ
    1305
    Rating
    4.50 / 5
    Context
    256K
    Price
    $0.8/M tokens
    License
    Open Weights
  12. 12

    Pixtral Large 2Mistral AI

    Mistral's multimodal flagship.

    IQ
    1300
    Rating
    4.50 / 5
    Context
    256K
    Price
    $2/M tokens
    License
    Open Weights
  13. 13

    MarbleWorld Labs

    Prompt in, walkable 3D world out.

    IQ
    1284
    Rating
    4.50 / 5
    Context
    ,
    Price
    not published
    License
    Closed
  14. 14

    Marengo 2.7Twelve Labs

    The video understanding model for search.

    IQ
    1276
    Rating
    4.60 / 5
    Context
    2h video
    Price
    $1.5/M tokens
    License
    Closed
  15. 15

    π0.5Physical Intelligence

    One brain, many robots, real homes.

    IQ
    1268
    Rating
    4.50 / 5
    Context
    ,
    Price
    not published
    License
    Open Weights
  16. 16

    EmberGoodfire

    See inside any open LLM.

    IQ
    1244
    Rating
    4.50 / 5
    Context
    ,
    Price
    not published
    License
    Closed

More rankings