2026-08-01에 감지된 AI 모델 가격·컨텍스트·종료일 변경 전체입니다.
| 변경 | 모델 | 내용 |
|---|---|---|
| 종료일 변경 | z-ai/glm-4.5v | 2026-12-31 -> - |
| 가격 변경 | deepseek/deepseek-v4-flash-0731 | 출력 -35.7% · 캐시 읽기 +542.9% · 입력 -35.7% |
| 가격 변경 | z-ai/glm-5.2 | 출력 -41.4% · 캐시 읽기 -41.4% · 입력 -41.4% |
| 가격 변경 | moonshotai/kimi-k2.6 | 출력 +37.5% · 캐시 읽기 +101.6% · 입력 +1.9% |
| 가격 변경 | openai/gpt-oss-20b | 출력 -7.1% · 캐시 읽기 과금 시작 |
| 가격 변경 | qwen/qwen3-coder-30b-a3b-instruct | 출력 +3.7% |
| 가격 변경 | thinkingmachines/inkling-small | 출력 -16.7% · 캐시 읽기 -13.8% · 입력 -13.8% |
| 가격 변경 | z-ai/glm-5.2 | 출력 -36.0% · 캐시 읽기 -36.0% · 입력 -36.0% |
| 가격 변경 | mistralai/mistral-small-3.2-24b-instruct | 출력 -33.3% · 캐시 읽기 과금 중단 · 입력 -25.0% |
| 가격 변경 | moonshotai/kimi-k2.6 | 출력 -38.0% · 캐시 읽기 -38.0% · 입력 -38.0% |
| 가격 변경 | nvidia/nemotron-3-nano-30b-a3b | 캐시 읽기 +20.0% |
| 가격 변경 | openai/gpt-oss-20b | 출력 +7.7% · 캐시 읽기 과금 중단 |
| 가격 변경 | poolside/laguna-s-2.1 | 출력 -10.0% · 캐시 읽기 -10.0% · 입력 -10.0% |
| 가격 변경 | qwen/qwen3-235b-a22b-thinking-2507 | 출력 -23.3% · 입력 -23.3% |
| 가격 변경 | qwen/qwen3-vl-30b-a3b-instruct | 출력 -13.3% · 입력 -13.3% |
| 가격 변경 | z-ai/glm-5.2 | 출력 +15.9% · 캐시 읽기 +15.9% · 입력 +15.9% |
| 신규 모델 | ~deepseek/deepseek-v4-flash-latest | 컨텍스트 1,048,576 |
| 신규 모델 | deepseek/deepseek-v4-flash-0731 | 컨텍스트 1,048,576 |
| 신규 모델 | thinkingmachines/inkling-small | 컨텍스트 524,288 |
| 모델 삭제 | mistralai/devstral-2512 | |
| 모델 삭제 | openai/gpt-5.1-chat | |
| 당사자 발언 | Tibo @thsottiaux | @RyanEls4 Codex || @RyanEls4 Codex |
| 당사자 발언 | Tibo @thsottiaux | @_chenglou 나는 그 팀의 일원이었어. 기본적으로 ChatGPT이 출시되기 1년 전의 ChatGPT이었지. LMChat이라고 불렸고 그다음에는 또 다른 코드명으로 불렸어. Google은 그걸 출시하기엔 너무 불안해했고 DeepMind는 Google을 뒤흔들 수 있는 제품을 출시하지 못하도록 막혀 있었어. 나는 이 일에 대해 많이 생각해. || @_chenglou I was part of that team. Basically ChatGPT one year before it came out. Called LMChat and then another codename. Google was too nervous to release it and DeepMind was blocked from shipping |
| RELEASE | b10219 | |
| OUTAGE | Incident with Copilot AI Model Providers | |
| 파라미터 변경 | deepseek/deepseek-v4-flash-0731 | 추가 top_a |
| 당사자 발언 | Elon Musk @elonmusk | 속도 & 비용을 고려할 때 Grok 4.5는 Pareto #1입니다 https://t.co/MoIM9hpHWr || Grok 4.5 is Pareto #1 when considering speed & cost https://t.co/MoIM9hpHWr |
| 당사자 발언 | Tibo @thsottiaux | @NanneWielinga @gdb 사실 모두입니다. 구독은 토큰이 매우 넉넉합니다. || @NanneWielinga @gdb All of them actually. The subscription is very generous in tokens. |
| RELEASE | b10218 | |
| ANNOUNCEMENT | Gemini 2.5 Pro and Gemini 3 Flash deprecated | |
| OUTAGE | Degraded availability GPT 5.6 Luna | |
| ANNOUNCEMENT | Ten advances in mathematics and theoretical computer science | |
| 파라미터 변경 | sao10k/l3-lunaris-8b | 추가 logprobs, top_logprobs |
| 파라미터 변경 | thinkingmachines/inkling-small | 추가 logprobs, tool_choice, tools, top_logprobs |
| CATALOG_VARIANT_DELISTED | ||
| 파라미터 변경 | poolside/laguna-s-2.1 | 삭제 stop |
| 컨텍스트 변경 | thedrummer/unslopnemo-12b | 32768 -> 1024000 |
| 파라미터 변경 | thedrummer/unslopnemo-12b | 추가 logit_bias, top_k |