Tinker cuts prices up to 70% after RL efficiency gains; GLM-5.3-Flash and DeepSeek-v4.1-Flash go live
ZhongRuiqi · x · 2026-10-10
Tinker, Thinking Machines' fine-tuning platform, announced price cuts of up to 70%, passing on efficiency gains from significant improvements that support users scaling up long-context RL training.
Two new cost-efficient models for long-context work, GLM-5.3-Flash and DeepSeek-v4.1-Flash, are also now live on the platform.
Related event: Tinker Cuts Prices by Up to 70% After Efficiency Upgrades(3 posts)→
More from Models
- After days of testing, dev says Sol 6.1 and Astra 6 fail at agentic work, returns to Opus 5.5 — NERDDISCO · 2026-10-11
- Qwen rumored to open-source both Qwen4-27 and Qwen4-max in a simultaneous release — HildeKuehne · 2026-10-11
- Ruff author Charlie Marsh plans a thread collecting model-release chart crimes — charliermarsh · 2026-10-11
- Leaked: OpenAI reportedly has internal models like 6.1 Astra ahead of Anthropic — haider1 · 2026-10-11
- Hugging Face downloads for top open LLMs drop ~30%, likely a counting change — natolambert · 2026-10-11
- Drex 1.5 lands on OpenRouter, billed as the fastest decision model available — Div_pradeep · 2026-10-10