400
changes shipped in the last two months that nobody announced. Prices moved, context windows shrank, capabilities disappeared — no email, no changelog entry, no error at runtime.
222
price changes
you pay more for the same call
226
capacity cuts
prompts truncate, no error
87
capabilities removed
a feature quietly went away
259
announced retirements
these ones you were told about
Evidence behind the numbers: 74 models confirmed against the provider's own API, 123 with dates taken from the provider's own documentation (extracted by a model, so labelled inferred), the remaining 2518 from third-party catalogs only. Every model page says which of the three it rests on.
Biggest silent changes
| Impact | Provider | What changed | Detected |
|---|---|---|---|
| truncation | gemini | -99% ctx · gemini/gemini-2.5-pro-preview-tts: context_tokens cut 1,048,576 to 8,192 | 2026-09-16 |
| truncation | openrouter | -98% max out · openrouter/moonshotai/kimi-k3:batch: max_output_tokens cut 943,718 to 16,384 | 2026-09-22 |
| truncation | bedrock_converse | -97% max out · zai.glm-4.7-flash: max_output_tokens cut 128,000 to 4,000 +1 alias | 2026-09-21 |
| cost | deepseek | 31.35x price · deepseek/deepseek-v4.1-flash: cache_read_price_per_mtok $0.0010/Mtok to $0.03/Mtok | 2026-09-27 |
| cost | ~deepseek | 31.35x price · ~deepseek/deepseek-flash-latest: cache_read_price_per_mtok $0.0010/Mtok to $0.03/Mtok | 2026-09-27 |
| cost | openrouter | 31.35x price · openrouter/deepseek/deepseek-v4.1-flash: cache_read_price_per_mtok $0.0010/Mtok to $0.03/Mtok | 2026-09-28 |
| truncation | thedrummer | -97% max out · thedrummer/unslopnemo-12b: max_output_tokens cut 1,024,000 to 32,768 inferred | 2026-08-24 |
| truncation | openai | -97% max out · openai/gpt-4.1-nano: max_output_tokens cut 942,818 to 32,768 | 2026-08-30 |
| truncation | nvidia | -96% max out · nvidia/nemotron-3-ultra-550b-a55b: max_output_tokens cut 461,059 to 16,384 | 2026-08-28 |
| cost | openrouter | 28.12x price · openrouter/deepseek/deepseek-v4-pro-0813: cache_read_price_per_mtok $0.0088/Mtok to $0.25/Mtok | 2026-09-27 |
| cost | deepseek | 27.79x price · deepseek/deepseek-v4-pro-0813: cache_read_price_per_mtok $0.0088/Mtok to $0.24/Mtok | 2026-09-27 |
| truncation | azure_ai | -96% max out · azure_ai/grok-4.3: max_output_tokens cut 200,000 to 8,192 | 2026-09-28 |
| cost | ~deepseek | 22.14x price · ~deepseek/deepseek-pro-latest: cache_read_price_per_mtok $0.0078/Mtok to $0.17/Mtok | 2026-09-27 |
| truncation | bedrock_converse | -95% max out · deepseek.v3.2: max_output_tokens cut 163,840 to 8,000 inferred | 2026-09-21 |
| cost | deepseek | 20.00x price · deepseek/deepseek-v4.1-flash: cache_read_price_per_mtok $0.0030/Mtok to $0.06/Mtok | 2026-09-25 |
| cost | ~deepseek | 20.00x price · ~deepseek/deepseek-flash-latest: cache_read_price_per_mtok $0.0010/Mtok to $0.02/Mtok | 2026-09-29 |
| cost | ~deepseek | 19.64x price · ~deepseek/deepseek-pro-latest: cache_read_price_per_mtok $0.01/Mtok to $0.25/Mtok | 2026-09-23 |
| cost | qwen | 16.87x price · qwen/qwen3.8-27b: input_price_per_mtok $0.02/Mtok to $0.42/Mtok | 2026-09-30 |
All 400 changes → · Check your own models →
Announced retirements
The part providers do tell you about. Included for completeness — it is the smaller half of the problem.
| Model | Vendor | Sunset | Remaining | Severity |
|---|---|---|---|---|
| google/gemini-omni-flash-preview | T-0d | critical | ||
| bedrock/amazon.nova-canvas | bedrock | T-0d | critical | |
| amazon/nova-canvas | amazon | T-0d | critical | |
| openai/gpt-5.4-cyber | openai | T-1d | critical | |
| azure/mai-image-2.5 | azure | T-1d | critical | |
| azure/mai-image-2.5-flash | azure | T-1d | critical | |
| azure/mai-image-2.5-pro | azure | T-1d | critical | |
| alibaba/qwen3-coder-30b-a3b-instruct | alibaba | T-1d | critical | |
| mistral/pixtral-12b-2409 | mistral | T-1d | critical | |
| google/gemini-2.5-flash-image | T-2d | critical |