~ same prompt, different LLM — every model gets the same brief in Blender, you judge the render ~
OpenAI / GPT 6 Astra / thinking:max has entered the arena NEW!
ARCHIVEDRetired from the leaderboard and voting — smoke-test scene — capability sanity check, not part of the benchmark · 2026-07-20. Preserved here for reference only.
Render Window — hero
hero render
camera + framing set up by the AI contestant
3D Viewport — red-ball User Persp
LOADING 3D VIEW
$0.20Cost
123sDuration
8Tool calls
9Turns
Blender 5.1.0MCP 1.22026-07-15
Tokens: 88419/3477Vision: Votes W-L-T: 0-0-0reasoning: effort mediumNO DONE
B-BENCH — TRANSCRIPT Anthropic / Opus 4.8 / thinking:medium · r1 · 2026-07-15
RUN LOG — red-ball · Anthropic / Opus 4.8 / thinking:medium · r1 · ok
SYSTEM PROMPTthe control — every model on this task got these instructions
PROMPTblender-bench v6HARNESSraw-v1
SHOW FULL PROMPT · 1104 chars
You are participating in Blender-Bench, a public benchmark where AI models complete the same Blender task.

Your result will be compared head-to-head with another model. Anyone on the internet can vote, based only on the published visual outputs they can inspect: the interactive 3D scene and rendered presentation.

Complete the request as well as you can. The benchmark does not prescribe a workflow or visual style; use whichever Blender techniques you consider appropriate. Make the important requirements evident in the published result, save a complete scene, and present the work clearly in every requested output.

## Harness Instructions

You are operating Blender through the provided tools. The scene starts empty. Build what the user's prompt asks for in the current scene. Work step by step and verify your own work. Before you finish, set up a camera and lighting to frame a single, well-composed presentation image of your result — position and aim the camera, and add or adjust lights so the subject reads clearly. When you are finished, reply with the single word DONE and no tool calls.
--TASK red-ball: ## Task Complete the following Blender task: <task> A red sphere of 1 meter diameter resting on a white ground plane — touching it exactly, neither floating nor sunk. Smooth shading, a plain red material on the ball, plain white on the floor. The floor is the sphere, the color, and the contact. The ceiling is presentation: a soft believable shadow under the ball, a hint of glossiness that catc…
49:35ai>I'll build this scene step by step. Let me start by setting up the sphere, floor, materials, lighting, and camera.
49:35>>execute_blender_code({"code":"\nimport bpy\n\n# Clean scene\nbpy.ops.object.select_all(action='SELECT')\nbpy.ops.object.delete(use_global=False)\nfor block in list(bpy.data.meshes): bpy.data.meshes.remove(block)\nfor block in list(bpy.data.materials): bpy.data.materials.remove(block)\n\n# Ground plane\nbpy.ops.mesh.primitive_plane_add(size=20, location=(0,0,0))\nfloor = bpy.context.active_object\nfloor.name = \"Floor\…)
49:35<<Code executed successfully: Sphere z: 0.5 radius: 0.5 bottom: 0.0
49:43ai>Now let me create the materials.
49:43>>execute_blender_code({"code":"\nimport bpy\n\nsphere = bpy.data.objects[\"RedBall\"]\nfloor = bpy.data.objects[\"Floor\"]\n\n# Red material with a hint of glossiness\nred = bpy.data.materials.new(\"RedMat\")\nred.use_nodes = True\nbsdf = red.node_tree.nodes[\"Principled BSDF\"]\nbsdf.inputs[\"Base Color\"].default_value = (0.8, 0.03, 0.03, 1.0)\nbsdf.inputs[\"Roughness\"].default_value = 0.25 # glossy hint\n# Specula…)
49:43<<Code executed successfully: Materials assigned
49:54ai>Now the camera, lighting, and render settings.
49:54>>execute_blender_code({"code":"\nimport bpy\nfrom mathutils import Vector\nimport math\n\nscene = bpy.context.scene\n\n# Camera\ncam_data = bpy.data.cameras.new(\"Camera\")\ncam = bpy.data.objects.new(\"Camera\", cam_data)\nscene.collection.objects.link(cam)\ncam.location = (2.8, -3.2, 1.8)\n# Aim at the ball center\ntarget = Vector((0,0,0.5))\ndirection = target - cam.location\ncam.rotation_euler = direction.to_track_…)
49:55<<Code executed successfully: Camera and lights set
49:57>>get_viewport_screenshot({"user_prompt":"check the scene composition"})
49:58<<
49:58imgimage delivered — get_viewport_screenshot
50:04ai>Scene is set up correctly. Let me render to verify the final presentation.
50:04>>execute_blender_code({"code":"\nimport bpy, os\nout = os.path.join(bpy.app.tempdir, \"render_check.png\")\nbpy.context.scene.render.filepath = out\nbpy.ops.render.render(write_still=True)\n# Save the blend\nblend_path = os.path.join(os.path.expanduser(\"~\"), \"red_sphere_scene.blend\")\nbpy.ops.wm.save_as_mainfile(filepath=blend_path)\nprint(\"Rendered to\", out)\nprint(\"Saved to\", blend_path)\n","user_prompt":"ren…)
50:17<<Code executed successfully: Info: Saved as "red_sphere_scene.blend" Rendered to <home>\AppData\Local\Temp\blender_a90352\render_check.png Saved to <home>\red_sphere_scene.blend
50:24ai>Let me view the render to verify the final result.
50:24>>execute_blender_code({"code":"\nimport bpy\n# Switch a viewport to rendered mode to preview\nfor area in bpy.context.screen.areas:\n if area.type == 'VIEW_3D':\n for space in area.spaces:\n if space.type == 'VIEW_3D':\n space.shading.type = 'RENDERED'\n # align view to camera\n for region in area.regions:\n if region.type == 'WINDOW':\n …)
50:24<<Code executed successfully: done
50:28>>get_viewport_screenshot({"user_prompt":"preview final rendered composition"})
50:32<<
50:32imgimage delivered — get_viewport_screenshot
50:44ai>The composition looks clean and the ball rests exactly on the floor with a soft shadow. The red looks a bit washed out though — let me deepen it and reduce the ambient wash for a richer red.
50:44>>execute_blender_code({"code":"\nimport bpy\n\nred = bpy.data.materials[\"RedMat\"]\nbsdf = red.node_tree.nodes[\"Principled BSDF\"]\nbsdf.inputs[\"Base Color\"].default_value = (0.65, 0.02, 0.02, 1.0)\nbsdf.inputs[\"Roughness\"].default_value = 0.2\n\n# Reduce world wash so red reads richer\nw = bpy.context.scene.world\nw.node_tree.nodes[\"Background\"].inputs[\"Strength\"].default_value = 0.25\n\n# Boost key a touch …)
51:18<<Code executed successfully: Info: Saved "red_sphere_scene.blend" done <home>\AppData\Local\Temp\blender_a90352\render_check2.png
51:21--status: budget_exhausted
>
FULL CALL INSPECTOR
OPEN RAW JSONL ↗
0 SHOWN
LOADING COMPLETE TRANSCRIPT…
Votes cast on this bench: 000482 · 157 runs · 25 contestants
Best viewed with Netscape Navigator 4.07 at 800×600 · real runs, live ratings