Top AI Product

What Is It

Z.AI (Zhipu AI) shipped GLM-5.1 — a 754B-parameter MoE model with 40B active parameters, open-sourced under MIT. It’s the first Chinese model to hit #1 on SWE-Bench Pro.

Score: 58.4. Ahead of GPT-5.4 (57.7), Claude Opus 4.6 (57.3), and Gemini 3.1 Pro (54.2). The entire model was trained on 100,000 Huawei Ascend 910B chips. Not a single Nvidia GPU touched this thing.

Why It Matters

The real story isn’t just the benchmark — it’s what it’s built on. US export controls were supposed to keep Chinese AI a generation behind on compute. Z.AI just topped the hardest coding benchmark in the industry on domestic silicon and gave the weights away for free.

Specs: 200K context window, 128K max output, supports up to 8 hours of autonomous execution for agentic workflows. API pricing at $1/M input and $3.20/M output — roughly 8x cheaper than Opus 4.6.

Benchmarks are self-reported and haven’t been independently verified yet. But if the numbers hold, open-source agentic coding just stopped being a Western monopoly.

You Might Also Like

• Siteline Just Dropped and it Might be the Analytics Tool the Agentic web Actually Needs
• Minimax m2 7 Scores 56 on swe pro and it Helped Build Itself
• Qwen 3 6 Plus vs Claude Opus 4 6 3x the Speed 1 17th the Price and the Benchmarks are Uncomfortably Close
• 26 Engineers 20m Arcee ai Trinity Large Thinking Scores Within 2 Points of Claude Opus
• Gpt 5 2 Theoretical Physics Discovery an ai Just Proved Physicists Wrong About Gluon Scattering