Skip navigation
Watch how Anthropic’s models have progressed over time.
0:00 / 1:08

Watch how Anthropic’s models have progressed over time.

Arena.ai

@arena244K followers

Visit on X
444 likesDownloadOriginal post

28,548 views

Watch how Anthropic’s models have progressed over time. Peter Gostev generated nine outputs from Opus 4.5 to Opus 5.5 (Max) using the same prompt. For more prompts to try on Arena, see the link below.

9 Comments

Comments from the original post on X (Twitter)

  • Arena.ai's profile picture

    @arena

    Check out @petergostev’s prompts at and try them yourself on Arena

    8 likes

  • Thiago S. Gusmao's profile picture

    @itsTh1ag0

    @AnthropicAI @petergostev Opus 5.5 is the "iPhone moment" from AI/SI

    3 likes

  • Ishan's profile picture

    @IshanJX

    @AnthropicAI @petergostev Improve like wine

    1 like

  • phon's profile picture

    @phon_t7

    @AnthropicAI @petergostev Good progress 🫡

  • miso's profile picture

    @miso_bits

    @AnthropicAI @petergostev This is pure gold, fr.

  • buildwithjay's profile picture

    @shipwithjay

    @AnthropicAI @petergostev Same prompt for every model is what makes this comparison mean anything. 3D is a good test too, since a wrong shape is obvious at a glance.

  • Nawaz Pasha's profile picture

    @nawaz_ai

    @AnthropicAI @petergostev Y was fable 5.2 not on this?

  • brb.'s profile picture

    @DOdeniyi9389

    @AnthropicAI @petergostev Love this. Do more of this please 🥺

  • liveta's profile picture

    @livetab_

    @AnthropicAI @petergostev Nine versions on one fixed prompt doubles as a regression baseline. Kept side by side, the outputs show what improved and also which instructions in the prompt newer models quietly stopped following.