What changed in AI
Every material change the connectors observe becomes an event with a source. Dates are when the change occurred (effective date when known); hover a date for the observation time. Older events load with a cursor.
6,520 events since 23 Jul 2026 · ordered by occurrence
14 Sept 2026 1
13 Sept 2026 49
OpenRouter changed pricing for DeepSeek V4 Flash 0423: $0.049 in / $0.098 out per 1M tokens → $0.04928 in / $0.09856 out per 1M tokens
$0.049 in / $0.098 out→$0.049 in / $0.099 outopenrouterllama.cpp: latest release at changed from 2026-09-13T05:17:02+00:00 to 2026-09-13T10:36:00+00:00
Latest release13 Sept 2026→13 Sept 2026githubllama.cpp: latest release tag changed from b10937 to b10941
Latest release tagb10937→b10941githubllama.cpp: latest version changed from b10937 to b10941
Latest versionb10937→b10941githubpytorch: latest release at changed from 2026-09-13T05:33:49+00:00 to 2026-09-13T07:39:42+00:00
Latest release13 Sept 2026→13 Sept 2026githubpytorch: latest release tag changed from trunk/c1c94458e8a382b3d84049f3c9f4961f59e8f609 to trunk/4d2bd99cd03cda454d98ff8c060c1353743c266f: flex_gemm…
Latest release tagtrunk/c1c94458e8a382b3d84049f3c9f4961f59e8f609→trunk/4d2bd99cd03cda454d98ff8c060c1353743c266f: flex_gemm: Inductor-owned QUACK autotuning (#196174)githubpytorch: latest version changed from trunk/c1c94458e8a382b3d84049f3c9f4961f59e8f609 to trunk/4d2bd99cd03cda454d98ff8c060c1353743c266f: flex_gemm…
Latest versiontrunk/c1c94458e8a382b3d84049f3c9f4961f59e8f609→trunk/4d2bd99cd03cda454d98ff8c060c1353743c266f: flex_gemm: Inductor-owned QUACK autotuning (#196174)githubOpenRouter changed pricing for DeepSeek V4 Flash 0423: $0.04928 in / $0.09856 out per 1M tokens → $0.049 in / $0.098 out per 1M tokens
$0.049 in / $0.099 out→$0.049 in / $0.098 outopenrouterOpenRouter changed pricing for DeepSeek V4 Flash 0423: $0.04956 in / $0.09912 out per 1M tokens → $0.04928 in / $0.09856 out per 1M tokens
$0.05 in / $0.099 out→$0.049 in / $0.099 outopenrouterTogether AI lists DeepSeek-V4.1-Flash at $0.3 in / $1.2 out per 1M tokens
together_pricingStep 3.5 Flash: hf repo changed from stepfun-ai/Step-3.5-Flash to stepfun-ai/step-3.5-flash
Hugging Face repostepfun-ai/Step-3.5-Flash→stepfun-ai/step-3.5-flashhuggingfacebartowski/Qwen3.8-27B-GGUF: last modified changed from 2026-09-11T19:46:27+00:00 to 2026-09-13T05:36:43+00:00
Last modified2026-09-11T19:46:27+00:00→2026-09-13T05:36:43+00:00huggingfacepytorch/pytorch released trunk/4d2bd99cd03cda454d98ff8c060c1353743c266f: flex_gemm: Inductor-owned QUACK autotuning (#196174)
githubOpenRouter changed pricing for DeepSeek V4 Flash 0423: $0.04984 in / $0.09968 out per 1M tokens → $0.04956 in / $0.09912 out per 1M tokens
$0.05 in / $0.10 out→$0.05 in / $0.099 outopenrouterKimi K2 0711: max output tokens changed from 100352 to 98304
Max output100.3K tokens→98.3K tokensopenrouterKimi K2 0905: max output tokens changed from 100352 to 98304
Max output100.3K tokens→98.3K tokensopenrouterKimi K2 Thinking: max output tokens changed from 100352 to 98304
Max output100.3K tokens→98.3K tokensopenrouterStep 3.5 Flash: hf repo changed from stepfun-ai/step-3.5-flash to stepfun-ai/Step-3.5-Flash
Hugging Face repostepfun-ai/step-3.5-flash→stepfun-ai/Step-3.5-Flashopenrouterexaone-4-5-33b: reasoning changed from false to true
ReasoningNo→Yesartificial_analysisQwen3 VL 30B A3B Instruct: reasoning changed from true to false
ReasoningYes→Noartificial_analysisQwen3 VL 32B Instruct: reasoning changed from true to false
ReasoningYes→Noartificial_analysisk-exaone: reasoning changed from false to true
ReasoningNo→Yesartificial_analysisminicpm5-1b: reasoning changed from false to true
ReasoningNo→Yesartificial_analysisgemma-4-E2B: reasoning changed from false to true
ReasoningNo→Yesartificial_analysisOpenRouter changed pricing for DeepSeek V4 Flash 0423: $0.0651 in / $0.1302 out per 1M tokens → $0.04984 in / $0.09968 out per 1M tokens
$0.065 in / $0.13 out→$0.05 in / $0.10 outopenrouterpytorch/pytorch released trunk/b8c2b362c474241d34ee1d2bc94213379334209c: [vllm hash update] update the pinned vllm hash (#196889)
githubpytorch/pytorch released viable/strict/1789295430: [vllm hash update] update the pinned vllm hash (#196889)
githubpytorch/pytorch released trunk/0468461462438fafdaaa819204279df108971489: [torchtitan hash update] update the pinned torchtitan hash (#196382)
githublitellm: latest release at changed from 2026-09-11T02:56:25+00:00 to 2026-09-13T04:56:38+00:00
Latest release11 Sept 2026→13 Sept 2026githublitellm: latest release tag changed from v1.102.0-dev.2 to v1.102.0-rc.1
Latest release tagv1.102.0-dev.2→v1.102.0-rc.1githublitellm: latest version changed from 1.102.0-dev.2 to 1.102.0-rc.1
Latest version1.102.0-dev.2→1.102.0-rc.1githublitellm: default branch changed from litellm_internal_staging to main
Default branchlitellm_internal_staging→maingithubllama.cpp: latest release at changed from 2026-09-12T23:11:26+00:00 to 2026-09-13T05:17:02+00:00
Latest release12 Sept 2026→13 Sept 2026githubllama.cpp: latest release tag changed from b10932 to b10937
Latest release tagb10932→b10937githubllama.cpp: latest version changed from b10932 to b10937
Latest versionb10932→b10937githubpytorch: latest release at changed from 2026-09-12T23:06:06+00:00 to 2026-09-13T05:33:49+00:00
Latest release12 Sept 2026→13 Sept 2026githubpytorch: latest release tag changed from trunk/609121c5bc919bb0141e62dc81273ab2c9785419: Add Nativ… to trunk/c1c94458e8a382b3d84049f3c9f4961f59e8f609
Latest release tagtrunk/609121c5bc919bb0141e62dc81273ab2c9785419: Add NativeAotStubs.h and NativeAotStubs.cpp to build→trunk/c1c94458e8a382b3d84049f3c9f4961f59e8f609githubpytorch: latest version changed from trunk/609121c5bc919bb0141e62dc81273ab2c9785419: Add Nativ… to trunk/c1c94458e8a382b3d84049f3c9f4961f59e8f609
Latest versiontrunk/609121c5bc919bb0141e62dc81273ab2c9785419: Add NativeAotStubs.h and NativeAotStubs.cpp to build→trunk/c1c94458e8a382b3d84049f3c9f4961f59e8f609githubOpenRouter changed pricing for DeepSeek V4 Flash 0423: $0.06538 in / $0.13076 out per 1M tokens → $0.0651 in / $0.1302 out per 1M tokens
$0.065 in / $0.131 out→$0.065 in / $0.13 outopenrouterR1: supported parameters changed from ["frequency_penalty", "include_reasoning", "max_tokens", … to ["frequency_penalty", "include_reasoning", "max_tokens", …
Supported parametersfrequency_penalty, include_reasoning, max_tokens, presence_penalty, reasoning, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_p→frequency_penalty, include_reasoning, max_tokens, presence_penalty, reasoning, repetition_penalty, response_format, seed, stop, temperature, tool_choice, tools, top_k, top_popenrouterGrok Voice Think Fast 1.0: regions changed from ["eu-west-1", "us-east-1", "us-saltlake-2"] to ["us-saltlake-2"]
Regionseu-west-1, us-east-1, us-saltlake-2→us-saltlake-2xaipytorch/pytorch released trunk/c1c94458e8a382b3d84049f3c9f4961f59e8f609
github
Feed keyed on occurred_at. Backfill = history imported when a source is first crawled (excluded — tick the box in the rail to include it). Times are UTC.