Ten models refreshed with 340 complete outputs
This refresh publishes 340 complete results across ten models, each with HTML, previews, a transcript and OpenRouter-reported usage. The models ran the current 35 prompts with the pi harness and browser tools available.
Explore ministral-14b-2512, step-3.5-flash, gemma-4-31b-it, qwen3-235b-a22b-2507, solar-pro4, glm-4.7-flash, nemotron-3-super-120b-a12b, seed-2.0-mini, gpt-4.1-nano, llama-3.3-70b-instruct.
Of 350 attempted runs, 10 ended with model or provider failures. Those 10 runs are excluded from this publication; their previously published data remains unchanged. Saved outcomes are retained for investigation.
Top builds in this update
OUTPUT ministral-14b-2512 Top-down alien shooter
Generated pi 0.87.1 · browser iteration · 469.2s · $0.0645 · 30k out
Full run + transcript →
OUTPUT step-3.5-flash Top-down alien shooter
Generated pi 0.87.1 · browser iteration · 129.9s · $0.0156 · 7.3k out
Full run + transcript →
OUTPUT gemma-4-31b-it Top-down alien shooter
Generated pi 0.87.1 · browser iteration · 287.2s · $0.0089 · 10k out
Full run + transcript →
OUTPUT · STOPPED BY AN API ERROR qwen3-235b-a22b-2507 Top-down alien shooter
Generated pi 0.87.1 · browser iteration · 281.0s · $0.0059 · 2.9k out
Full run + transcript →