~ same prompt, different LLM — every model gets the same brief in Blender, you judge the render ~
Google / Gemini 3.8 Flash / thinking:medium has entered the arena NEW!
LeaderboardGalleryVoteRunsTasks
RankContestantRatingVotesRunsDone %Success %Avg $Avg callsAvg time
1OpenAI / GPT 5.6 Sol / thinking:maxcompare →1794251100100$1.84—980s
2OpenAI / GPT 6 Astra / thinking:mediumcompare →17502510100$0.57—296s
3OpenAI / GPT 5.6 Lunacompare →15772510100$0.04—74s
4Google / Gemma 4 31b Itcompare →12881910100$0.00—173s
5DeepSeek / Deepseek V4 Flashcompare →12041710100$0.00—122s
provAnthropic / Sonnet 5 / thinking:lowcompare →1387110100$0.08—100s

Model comparison

each point is a model (One-shot, cg-still-life) — hover for exact values
AnthropicDeepSeekGoogleOpenAI unrated
11161307149916911883$0.001$0.007$0.053$0.414$3.24Bradley–Terry ratingOpenAI / GPT 5.6 Sol / thinking:maxOpenAI / GPT 6 Astra / thinking:mediumOpenAI / GPT 5.6 LunaGoogle / Gemma 4 31b ItDeepSeek / Deepseek V4 FlashAnthropic / Sonnet 5 / thinking:lowMean cost per task (USD, log scale)

By dimension

separate Bradley–Terry fits from lens rounds + chip verdicts
Lighting
provOpenAI / GPT 5.6 Sol / thinking:max1595
provOpenAI / GPT 5.6 Luna1405
Composition
no data yet
Modeling
no data yet
Materials
no data yet
Render
no data yet
Adherence
no data yet

Modes compared

each contestant's ratings across run modes (mixed ladder)
ContestantFullOne-shot3D-only
DeepSeek / Deepseek V4 Flash—1194—
Google / Gemma 4 31b It—1267—
Z.AI / GLM 5.2———
OpenAI / GPT 5.6 Luna14971531—
OpenAI / GPT 5.6 Luna / thinking:max1814——
OpenAI / GPT 5.6 Sol1713——
OpenAI / GPT 5.6 Sol / thinking:max—1750—
OpenAI / GPT 5.6 Terra1558——
OpenAI / GPT 6 Astra / thinking:medium—1710—
Moonshot AI / Kimi K3 / thinking:max———
Anthropic / Sonnet 51142——
Anthropic / Sonnet 5 / thinking:low—1323—
Vote A/BGalleryRunsbradley-terry fit over 56 decisive votes
Votes cast on this bench: 001671 · 163 runs · 26 contestants
Best viewed with Netscape Navigator 4.07 at 800×600 · real runs, live ratings