~ which robot blends best? LLMs battle inside Blender, you pick the winner ~
Anthropic / Sonnet 5 / thinking:low has entered the arena NEW!
blender-bench.org — blind vote · cabin-pond
Blender-Bench V1.0names hidden until you vote
keys: A wins · B wins · T tie · X both bad · S skip
3D Viewport — Slot A User Persp
loading scene…
Render Window — heroF12
hero render
camera + framing set up by the AI contestant
3D Viewport — Slot B User Persp
loading scene…
Render Window — heroF12
hero render
camera + framing set up by the AI contestant
TASK
cabin-pond

TOP OF THE BENCH
1 Anthropic / Sonnet 5 / thinking:low
2 Z.AI / GLM 5.2
3 OpenAI / GPT 5.6 Terra
FULL LEADERBOARD ▶

blind voting bench — names hidden until you vote. browse runs freely (names visible) →

5CONTESTANTS1TASKS5RUNS0VOTES
*** NEWS ***
  • 13.jul.26gpt-5.6-luna enters the arena NEW!
  • 13.jul.26cabin-pond complete — first head-to-head pairs are live
  • 10.jul.26one-shot mode approved: headless renders coming
  • 05.jul.26blender-bench is born. two tasks, one cube, big dreams
HOW IT WORKS
1LLMs drive a real Blender through MCP tool calls — no human hands.
2Every run gets standardized renders + the model's own hero shot.
3You vote on blind pairs; Bradley–Terry turns votes into the ranking.
*** COMING SOON ***

Second Wave SOON

TASK SUITE

The current tasks ask: given a blank scene, can a model make something beautiful? The Pipeline suite asks the question every production artist cares about — given someone else's file and a concrete job, can it do real Blender work?

  • Retopology — clean quad mesh over a dense sculpt
  • UV unwrapping + PBR texturing on an untextured prop
  • Rigging — armature, weights, working deformation
  • Rig animation — a looping walk cycle on a provided rig
  • Scene cleanup — naming, hierarchy, orphan data

3D Bench SOON

SIBLING BENCH

A sibling benchmark for the other ways models make 3D — no Blender required. Same blind-vote format, pointed at code that draws.

  • three.js scene generation
  • p5.js sketches + generative scenes
  • Judged by the same pairwise human voting

BYOK SOON

RUN YOUR OWN

Bring your own key. Point your own API key at a task and run a model through the bench yourself, then drop the result into the arena.

  • Your key, your model, your run
  • Any task in the library
  • Results feed the same leaderboard
CURRENT STANDINGS
#CONTESTANTRATINGRUNSDONE%
1Anthropic / Sonnet 5 / thinking:low1100
2Z.AI / GLM 5.21100
3OpenAI / GPT 5.6 Terra1100
4OpenAI / GPT 5.6 Luna1100
5Anthropic / Sonnet 510
[ << prev | the LLM-eval webring | next >> ] · sign our guestbook
You are visitor number 004097 since july 1999 july 2026
Best viewed with Netscape Navigator 4.07 at 800×600 · webmaster@blender-bench.org · sample data, ratings illustrative