Skip to content
AI Atlas
Changes

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.

10,000 events since 20 May 2026 · ordered by occurrence

16 Sept 2026 50

  • AnnouncementCompanyCohere

    Cohere: Cohere and Aleph Alpha sign agreement to become the first transatlantic sovereign AI solution

    cohere
  • New paperPaperarXiv:2510.22115

    New paper: arXiv:2510.22115

    huggingface
  • New paperPaperarXiv:2507.17634

    New paper: arXiv:2507.17634

    huggingface
  • New modelModelgranite-3.3-2b-instructIBM

    New model: granite-3.3-2b-instruct (IBM)

    huggingface
  • New artifactArtifactunsloth/gemma-4-12b-it-GGUFUnsloth

    New artifact: unsloth/gemma-4-12b-it-GGUF (Unsloth)

    huggingface
  • New artifactArtifactmlx-community/GLM-OCR-4bitMLX Community

    New artifact: mlx-community/GLM-OCR-4bit (MLX Community)

    huggingface
  • New modelModelinternlm2_5-7bInternLM (Shanghai AI Laboratory)

    New model: internlm2_5-7b (InternLM (Shanghai AI Laboratory))

    huggingface
  • New modelModelgranite-speech-3.3-2bIBM

    New model: granite-speech-3.3-2b (IBM)

    huggingface
  • New modelModelc4ai-command-r-plusCohere

    New model: c4ai-command-r-plus (Cohere)

    huggingface
  • New modelModelSeed-OSS-36B-Base-woSynByteDance

    New model: Seed-OSS-36B-Base-woSyn (ByteDance)

    huggingface
  • AnnouncementCompanyOpenAI

    OpenAI: Reimagining advertising with AI

    openai
  • Price changedModelMythoMax 13Bgryphe

    OpenRouter changed pricing for MythoMax 13B: $0.06 in / $0.06 out per 1M tokens → $0.08 in / $0.11 out per 1M tokens

    $0.06 in / $0.06 out$0.08 in / $0.11 outopenrouter
  • Property changedModelUnslopNemo 12BTheDrummer

    UnslopNemo 12B: tool calling changed from true to false

    Tool callingYesNoopenrouter
  • Property changedModelUnslopNemo 12BTheDrummer

    UnslopNemo 12B: supported parameters changed from ["frequency_penalty", "logit_bias", "logprobs", "max_toke… to ["frequency_penalty", "logit_bias", "logprobs", "max_toke…

    Supported parametersfrequency_penalty, logit_bias, logprobs, max_tokens, presence_penalty, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_k, top_logprobs, top_pfrequency_penalty, logit_bias, logprobs, max_tokens, presence_penalty, repetition_penalty, seed, stop, structured_outputs, temperature, top_k, top_logprobs, top_popenrouter
  • Property changedFrameworkqdrantQdrant

    qdrant: pypi release at changed from 2026-08-04T14:32:55+00:00 to 2026-09-16T06:43:12+00:00

    Pypi release at4 Aug 202616 Sept 2026pypi
  • Property changedFrameworkqdrantQdrant

    qdrant: pypi version changed from 1.19.0 to 1.19.1

    Pypi version1.19.01.19.1pypi
  • Price changedModelQwen3.7 MaxQwen

    Together AI changed pricing for Qwen3.7 Max: $2 in / $6 out per 1M tokens → $2.5 in / $7.5 out per 1M tokens

    $2 in / $6 out$2.5 in / $7.5 outtogether_pricing
  • Price changedModelGLM 5.3 FlashZ.ai (Zhipu AI)

    OpenRouter changed pricing for GLM 5.3 Flash: $0.1 in / $0.3333 out per 1M tokens → $0.09 in / $0.3 out per 1M tokens

    $0.10 in / $0.333 out$0.09 in / $0.30 outopenrouter
  • AnnouncementCompanyCohere

    Cohere: Cohere and OpenText partner to bring trusted agentic AI to governments and regulated industries

    cohere
  • Property changedFrameworklangfuseLangfuse

    langfuse: latest release at changed from 2026-09-15T09:00:35+00:00 to 2026-09-16T09:06:14+00:00

    Latest release15 Sept 202616 Sept 2026github
  • Property changedFrameworklangfuseLangfuse

    langfuse: latest release tag changed from v4.36.1 to v4.37.0

    Latest release tagv4.36.1v4.37.0github
  • Version releasedFrameworklangfuseLangfuse

    langfuse: latest version changed from 4.36.1 to 4.37.0

    Latest version4.36.14.37.0github
  • Property changedFrameworklancedbLanceDB

    lancedb: latest release at changed from 2026-09-15T21:06:11+00:00 to 2026-09-16T09:26:40+00:00

    Latest release15 Sept 202616 Sept 2026github
  • Property changedFrameworklancedbLanceDB

    lancedb: latest release tag changed from LanceDB v0.39.0-beta.9 to LanceDB v0.39.0-beta.10

    Latest release tagLanceDB v0.39.0-beta.9LanceDB v0.39.0-beta.10github
  • Version releasedFrameworklancedbLanceDB

    lancedb: latest version changed from LanceDB v0.39.0-beta.9 to LanceDB v0.39.0-beta.10

    Latest versionLanceDB v0.39.0-beta.9LanceDB v0.39.0-beta.10github
  • Property changedAgentcodexOpenAI

    codex: latest release at changed from 2026-09-16T01:38:01+00:00 to 2026-09-16T04:23:12+00:00

    Latest release16 Sept 202616 Sept 2026github
  • Property changedAgentcodexOpenAI

    codex: latest release tag changed from 0.155.0-alpha.9 to 0.155.0-alpha.10

    Latest release tag0.155.0-alpha.90.155.0-alpha.10github
  • Version releasedAgentcodexOpenAI

    codex: latest version changed from 0.155.0-alpha.9 to 0.155.0-alpha.10

    Latest version0.155.0-alpha.90.155.0-alpha.10github
  • Property changedFrameworkllama.cppggml.ai

    llama.cpp: latest release at changed from 2026-09-15T23:46:09+00:00 to 2026-09-16T09:06:21+00:00

    Latest release15 Sept 202616 Sept 2026github
  • Property changedFrameworkllama.cppggml.ai

    llama.cpp: latest release tag changed from b10991 to b10997

    Latest release tagb10991b10997github
  • Version releasedFrameworkllama.cppggml.ai

    llama.cpp: latest version changed from b10991 to b10997

    Latest versionb10991b10997github
  • Property changedFrameworkflash-attentionDao AI Lab

    flash-attention: latest release at changed from 2026-09-09T08:20:40+00:00 to 2026-09-16T08:21:28+00:00

    Latest release9 Sept 202616 Sept 2026github
  • Property changedFrameworkflash-attentionDao AI Lab

    flash-attention: latest release tag changed from fa4-v4.0.0.beta30 to fa4-v4.0.0.beta31

    Latest release tagfa4-v4.0.0.beta30fa4-v4.0.0.beta31github
  • Version releasedFrameworkflash-attentionDao AI Lab

    flash-attention: latest version changed from fa4-v4.0.0.beta30 to fa4-v4.0.0.beta31

    Latest versionfa4-v4.0.0.beta30fa4-v4.0.0.beta31github
  • Property changedFrameworkpytorchPyTorch Foundation

    pytorch: latest release at changed from 2026-09-16T03:35:08+00:00 to 2026-09-16T06:11:27+00:00

    Latest release16 Sept 202616 Sept 2026github
  • Property changedFrameworkpytorchPyTorch Foundation

    pytorch: latest release tag changed from trunk/499da496df69848dd2921da2c319124f76f0485f to viable/strict/1789554516

    Latest release tagtrunk/499da496df69848dd2921da2c319124f76f0485fviable/strict/1789554516github
  • Version releasedFrameworkpytorchPyTorch Foundation

    pytorch: latest version changed from trunk/499da496df69848dd2921da2c319124f76f0485f to viable/strict/1789554516

    Latest versiontrunk/499da496df69848dd2921da2c319124f76f0485fviable/strict/1789554516github
  • ReleaseAgentcodexOpenAI

    openai/codex released rusty-v8-v152.2.0

    github
  • ReleaseFrameworklancedbLanceDB

    lancedb/lancedb released LanceDB v0.39.0-beta.10

    github
  • ReleaseFrameworkllama.cppggml.ai

    ggml-org/llama.cpp released b10997

    github
  • ReleaseFrameworklangfuseLangfuse

    langfuse/langfuse released v4.37.0

    github
  • New modelModelOpenELM-1_1BApple

    New model: OpenELM-1_1B (Apple)

    huggingface
  • ReleaseFrameworkllama.cppggml.ai

    ggml-org/llama.cpp released b10996

    github
  • ReleaseFrameworkvllmvLLM project

    vllm-project/vllm released proto-v0.2.0: vllm-proto 0.2.0

    github
  • ReleaseFrameworkflash-attentionDao AI Lab

    Dao-AILab/flash-attention released fa4-v4.0.0.beta31

    github
  • ReleaseFrameworkllama.cppggml.ai

    ggml-org/llama.cpp released b10995

    github
  • ReleaseFrameworkllama.cppggml.ai

    ggml-org/llama.cpp released b10994

    github
  • Price changedModelQwen3 VL 30B A3B InstructQwen

    OpenRouter changed pricing for Qwen3 VL 30B A3B Instruct: $0.13 in / $0.52 out per 1M tokens → $0.15 in / $0.6 out per 1M tokens

    $0.13 in / $0.52 out$0.15 in / $0.60 outopenrouter
  • Max output changedModelQwen3 VL 30B A3B InstructQwen

    Qwen3 VL 30B A3B Instruct: max output tokens changed from 32768 to 16384

    Max output32.8K tokens16.4K tokensopenrouter
  • ReleaseAgentcodexOpenAI

    openai/codex released rust-v0.155.0-alpha.2.5

    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.