Near-frontier agentic coding performance from a Qwen3.8-27B post-train that quantizes to…
This is a dev post classified by AGK Intelligence as AI dev tools (a tool drop), kept by the Dev Radar because it carries real work, not commentary.
Near-frontier agentic coding performance from a Qwen3.8-27B post-train that quantizes to ~17GB. 👀 Meaning this can run on your 24GB RTX 3090. BAAI's AREX-2 took Qwen3.8-27B and post-trained it to keep improving a solution over multiple attempts, and the improvements are crazy. MLE-bench Lite: 🔥 AREX-2 27B → 81.8 GPT-5.6 Sol → 72.7 Kimi-K3 2.8T → 72.7 Kimi-K2.6 1T → 66.7 DeepSeek V4 Pro 1.6T → 54.5 Frontier-CS: GPT-5.6 Sol → 76.4 Claude Opus 4.8 → 74.5 🔥 AREX-2 27B → 70.7 Gemini-3.1-Pro → 68.9 Kimi-K2.7-Code 1T → 54.7 GLM-5.3-Flash 320B → 50.4 And BAAI actually did the important ablation.
Posted by David Hendrickson (12.6k followers) 3 days ago · 118 likes · 6.9k views · view the original post on X. Kept by the Dev Radar as AI dev tools.
More dev work like this
- 长PDF要是拆成单页再往一块拼，跨页表格和阅读顺序十有八九会乱套，这事你八成也烦过。 — @bkdgiffug
- Local models on Apple Silicon finally feel like a product, not a weekend project. — @PixelRainbowNFT
- GPT 6.1 Terra max vs GPT 6.1 Sol max — @HarshithLucky3
- Asked at the Meetup: "Does sbx run mistral work?" — @ajeetsraina
- Most agents freeze once deployed. — @GoogleCloudTech
- Your $250 GPU can now run a 35B coding model at 50 tokens/sec. — @0x0SojalSec
- Next week we're showing how to build self improving products. — @spenserskates
- Seedance 2.5 US Draft mode is now available on fal — @fal
Every post is read and classified by AGK Intelligence: what it is, which market it belongs to, and whether the link is a real tool. 31.2k posts from 5k X accounts over the last 21 days, 8.8k tools, 12 markets. New posts are collected several times an hour and the full ranking is recomputed every hour — last update 2026-10-30 22:12 UTC. Full method.