Every AI model price, context and end-of-life change detected on 2026-08-25.
| Change | Model | Detail |
|---|---|---|
| END OF LIFE | moonshotai/kimi-k2.5 | shuts down 2026-08-31 · 6 days left |
| END OF LIFE | dots-studio/dots-3-note-preview:free | shuts down 2026-09-30 · 36 days left |
| END OF LIFE | z-ai/glm-4.5 | shuts down 2026-12-31 · 128 days left |
| END OF LIFE MOVED | z-ai/glm-4.5v | 2026-12-31 -> - |
| PRICE CHANGE | deepseek/deepseek-v4-flash | output +47.5% · cache read +47.5% · input +47.5% |
| PRICE CHANGE | deepseek/deepseek-v4-flash-0731 | output -39.2% · cache read -39.2% · input -39.2% |
| PRICE CHANGE | deepseek/deepseek-v4-flash-vision-exp | overrides now charged |
| PRICE CHANGE | deepseek/deepseek-v4-pro | output +9.7% · cache read +9.7% · input +9.7% |
| PRICE CHANGE | qwen/qwen3.8-27b | output -15.0% |
| PRICE CHANGE | thinkingmachines/inkling | cache read -5.9% · input -5.0% |
| PRICE CHANGE | z-ai/glm-5.1 | output +30.4% · cache read +30.4% · input +30.4% |
| PRICE CHANGE | z-ai/glm-5.2 | output +23.2% · cache read +14.4% · input +23.2% |
| PRICE CHANGE | ~deepseek/deepseek-v4-flash-latest | output +25.0% · input -12.5% |
| PRICE CHANGE | ~moonshotai/kimi-latest | output +7.7% · input +7.7% |
| NEW MODEL | minimax/minimax-m2.7:free | context 196,608 |
| NEW MODEL | minimax/minimax-m3:free | context 1,048,576 |
| MODEL REMOVED | qwen/qwen-plus-2025-07-28:thinking | |
| FROM THE SOURCE | OpenAI @OpenAI | Jalapeño means faster ChatGPT responses, more responsive Codex sessions and agents, and reliable access as demand continues to grow. https://t.co/QgBKRa3Xz6 |
| FROM THE SOURCE | OpenAI @OpenAI | Since announcing Jalapeño, our first custom inference chip, we’ve been testing it and the system around it. The results show a major advance: more intelligence from every watt and faster responses, delivering both higher throughput and lowe |
| OUTAGE | Issues logging into Claude.ai | |
| RELEASE | v2.1.245 | |
| RELEASE | v0.3.0 | |
| ANNOUNCEMENT | Push rules in rulesets now support path exceptions | |
| ANNOUNCEMENT | Your alt text passes automated checks. That doesn’t mean it’s any good. | |
| ANNOUNCEMENT | Quantization-Aware Healing: a compressed, 4-bit model that outperforms its full-precision original | |
| ANNOUNCEMENT | Wire It, Run It, Deploy It: AI Workflows in Gradio | |
| ANNOUNCEMENT | Granite 4.2 LLMs: How They're Built | |
| ANNOUNCEMENT | How Hugging Face Inference Endpoints, Jobs, and Buckets Power Search on Papers with Code | |
| ANNOUNCEMENT | Disrupting a new covert influence campaign from Russia | |
| ANNOUNCEMENT | Introducing the Admin plugin for ChatGPT Work and Codex | |
| ANNOUNCEMENT | Jalapeño’s first results show industry-leading speed and efficiency in AI inference | |
| ANNOUNCEMENT | The full stack behind abundant intelligence | |
| CONTEXT CHANGED | thinkingmachines/inkling-small:free | 262144 -> 1048576 |
| CONTEXT CHANGED | thinkingmachines/inkling:free | 262144 -> 1048576 |
| PARAMETERS CHANGED | x-ai/grok-4.6 | added stop, top_k |
| PARAMETERS CHANGED | ~x-ai/grok-latest | added stop, top_k |