Same 3D prompt one-shot across 8 models incl. Mistral Large 4
DesignRead the post
x.com · link re-checked 2026-10-16
DeepSeek V4 Pro
By DeepSeek. 4 results on 4 boards, copied as published. Different boards, different metrics: don't add them up.
License (as listed)MIT
Hosting regionNot confirmed
Boards4
Code
- 40.2%SWE-rebench
DeepSeek-V4 Pro [high] · Resolved rate % · board #14
SWE-rebench harness (agent rows are full products) · as of 2026-07-01 Stale ↗
Agentic / terminal
No result on the boards we captured
Websites
- 1447LMArena WebDev Arena
deepseek-v4-pro · Arena score · board #78
LMArena WebDev (agentic front-end tasks) · as of 2026-10-22 ↗
Design
- 1239Design Arena (All categories)
deepseek-v4-pro · Elo · board #62
Design Arena · as of 2026-10-24 ↗
Android
- 59.5%Android Bench 1.0 tasks (100 tasks)
Deepseek V4 Pro · Score %
not_stated · as of 2026-08-11 ↗
iOS
No mature public leaderboard
Available in
No plan row in our pricing data names this model version.
Use cases with this model
Spellbook of Skill: curated dev-skills guide site
WebsitesRead the post
dev.to · web source · link re-checked 2026-10-16
Cline Desktop + DeepSeek V4 Pro: "pointed it at an empty folder, and asked it to build an evaluation harness for an ML…"
CodeRead the post
x.com · link re-checked 2026-10-16
Same family: DeepSeek 14
"I built a 3D BMW M5 with AI for $0.00 … Model: DeepSeek V4.1 Flash - Tool: OpenCode - Total tokens: 76K - Cost: $0.00"
DesignRead the post
x.com · link re-checked 2026-10-16
An inventor's attic workshop as a full three.js scene, "all one-shot"
DesignRead the post
x.com · link re-checked 2026-10-16
Test: "Qwen 3.8 Max (max): 25.7 - DeepSeek V4.1 Flash (max): 21.7 - Kimi K3 (default): 23 - GLM-5.3 (max): 19 - Mistral Large 4 (high, its max): 15"
CodeRead the post
x.com · link re-checked 2026-10-16
Self-playing Flappy Bird and Missile Command games, each one-shot in a single HTML file, comparing 4 open models
WebsitesDesignRead the post
empiriolabs.ai · web source · link re-checked 2026-10-16
"$460 in tokens → ~$22K in bounties" from multi-model hackbots on open-weight models, #1 on a HackerOne leaderboard
CodeRead the post
x.com · link re-checked 2026-10-16
"I built a working AI agent in 1.5 hours for $0.49" (Pieni; DeepSeek among supported backends)
CodeRead the post
x.com · link re-checked 2026-10-16
Sources
lmarena.ai/leaderboard/webdev swe-rebench.com developer.android.com/bench designarena.ai/leaderboard