Same base model, 6x agentic jump: GLM-5.3 hits 28.3% on Terminal-Bench via post-training
bittingthembits · x · 2026-08-25
GLM-5.3 keeps the same base model as GLM-5.2, but scaled post-training — more environments, more diverse tasks, more compute — lifting Terminal-Bench 3.0 resolution from 4.6% to 28.3%, roughly a 6x jump. Terminal-Bench tests agents on complex tasks in real terminal environments.
The poster argues pretraining (learning language, code, patterns) is dominated by big labs with massive GPU budgets, while post-training (tasks, environments, feedback, rewards) is a far more open race — which Affine, a Bittensor subnet, has turned into an ongoing open competition.
More from Models
- Fine-tuning Models to Remove Slop: Practical Insights — ivan_bezdomny · 2026-08-25
- Startup's neural-operator model may be source of huge-context-window rumors — McDonaghMatthew · 2026-08-25
- Ex-NVIDIA scientist's startup launches physics AI model handling 5T data points per prompt — thesaraharminta · 2026-08-25
- OpenAI blog: Codex brought three open-weight models to high performance in two months — eliebakouch · 2026-08-25
- Users notice Codex usage limits appear to have been silently reset — kevinkern · 2026-08-25
- Arav Srinivas: On-device models critical for sensitive docs with SOTA OCR — AravSrinivas · 2026-08-25