{"assertions":[{"assertion_index":0,"assertion_text":"Anthropic News TechCrunch AI Yesterday the first team arrived.","citation_check":null,"claim_ids":[55277,61214,61215],"link_only_urls":[]},{"assertion_index":1,"assertion_text":"Anthropic News The question is what the badge buys.","citation_check":null,"claim_ids":[79434,79435,79436],"link_only_urls":[]},{"assertion_index":2,"assertion_text":"TechCrunch AI METR's record is the best answer to it: in March METR found multiple problems with the reasoning in Anthropic's sabotage risk report for Claude Opus 4.6, and in May it judged the evidence in Anthropic's R&D automation section too thin to establish low catastrophe risk.","citation_check":null,"claim_ids":[62516],"link_only_urls":[]},{"assertion_index":3,"assertion_text":"METR METR A lab inviting that reviewer inside pays a real cost.","citation_check":null,"claim_ids":[21612,21826,21832],"link_only_urls":[]},{"assertion_index":4,"assertion_text":"TechCrunch AI Latent Space TechCrunch AI Signing a standard is not the same as letting someone in, and of the three signers, Anthropic's onboarding is the only one described in the evidence I have.","citation_check":null,"claim_ids":[61559,62516,71218],"link_only_urls":[]},{"assertion_index":5,"assertion_text":"Latent Space None of this is a reason to doubt the commitment.","citation_check":null,"claim_ids":[59495],"link_only_urls":[]},{"assertion_index":6,"assertion_text":"METR What is new is standing, employee-level access with no vendor veto over publication, so make the published report a condition in your next renewal conversation rather than a reason to sign today.","citation_check":null,"claim_ids":[21645],"link_only_urls":[]},{"assertion_index":7,"assertion_text":"TechCrunch AI Latent Space Muse Spark Open looks like Meta making good on that promise, though nothing in the evidence here ties it to the 1.3 version.","citation_check":null,"claim_ids":[44762,44769,56122],"link_only_urls":[]},{"assertion_index":8,"assertion_text":"Hugging Face Blog Transformers Releases For an engineer deciding whether to self-host, which means running the model on your own servers, the two numbers answer different questions: 42 billion sets per-token compute, and 816 billion sets the memory bill.","citation_check":null,"claim_ids":[79484,79485,79486,79716,79717,79718],"link_only_urls":[]},{"assertion_index":9,"assertion_text":"Latent Space Price the hosted route alongside your own hardware before committing either way.","citation_check":null,"claim_ids":[24045],"link_only_urls":[]},{"assertion_index":10,"assertion_text":"OpenRouter Blog That mostly continues a pattern: private model routing in June and hard workspace spend caps in August were also Enterprise-only, though data-residency routing, which keeps requests inside the US or EU, reaches the Business tier as well.","citation_check":null,"claim_ids":[79282,79283,79284],"link_only_urls":[]},{"assertion_index":11,"assertion_text":"OpenRouter Blog OpenRouter Blog OpenRouter Blog The routing layer is cheap; the control layer is not.","citation_check":null,"claim_ids":[31070,43150,58935,58943],"link_only_urls":[]},{"assertion_index":12,"assertion_text":"OpenRouter Blog","citation_check":null,"claim_ids":[30832],"link_only_urls":[]},{"assertion_index":13,"assertion_text":"Mistral AI Those are tiny by language-model standards, and deliberately so.","citation_check":null,"claim_ids":[79577,79578,79579],"link_only_urls":[]},{"assertion_index":14,"assertion_text":"Hugging Face Blog Hugging Face Blog Mistral is joining a crowded small-model lane rather than opening one.","citation_check":null,"claim_ids":[59189,69566],"link_only_urls":[]},{"assertion_index":15,"assertion_text":"AI Engineer Run it on your own series before anyone's chart.","citation_check":null,"claim_ids":[37640],"link_only_urls":[]},{"assertion_index":16,"assertion_text":"- METR's first public document from inside Anthropic will show whether the redaction categories, which include customer data and legal privilege, stay narrow in practice. Anthropic News","citation_check":null,"claim_ids":[79434,79435,79436],"link_only_urls":[]},{"assertion_index":17,"assertion_text":"- OpenAI and xAI cosigned the AEF-1 evaluator standard; whether either follows Anthropic with an on-site team is the open question. Latent Space","citation_check":null,"claim_ids":[61559],"link_only_urls":[]},{"assertion_index":18,"assertion_text":"- Meta's Muse Spark Open now has native Transformers support, so independent runs on the 816 billion parameter weights should surface quickly. Transformers Releases","citation_check":null,"claim_ids":[79716,79717,79718],"link_only_urls":[]},{"assertion_index":19,"assertion_text":"- Mistral said it would release the Mistral Large 4 weights by the end of October, which is two weeks away. Mistral AI","citation_check":null,"claim_ids":[73723],"link_only_urls":[]},{"assertion_index":20,"assertion_text":"- OpenRouter's Enterprise-only Private Endpoints will test how many teams pay for gateway governance rather than route around it. OpenRouter Blog","citation_check":null,"claim_ids":[79282,79283,79284],"link_only_urls":[]},{"assertion_index":21,"assertion_text":"Anthropic News Muse Spark Open's parameter figures come from the launch post rather than independent runs, its link to the promised 1.3 release is my inference, the under-2 GB memory figure for small models comes from a conference talk rather than a Timestral measurement, and Timestral's release carries no accuracy comparison in the evidence I have.","citation_check":null,"claim_ids":[79434,79435,79436],"link_only_urls":[]}],"brief_date":"2026-10-16","bundle_id":"evidence-bundle-v2-5e620c227fa9b5aefccc3f7f2bdf0a45b2d6a9d66c2b305544e8c4c6e18d20f2","claims":[{"assertion_indexes":[0],"citation_authority_claim_id":55277,"citation_checks":[],"claim_id":55277,"claim_text":"Anthropic is in dialogue with METR, an independent AI evaluation organization, to conduct a third-party review including access to all transcripts and sampling access to the relevant models.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-09-03T06:30:31.464692+00:00","source_name":"Anthropic News","source_url":"https://www.anthropic.com/news/investigating-incidents-cybersecurity-evals"},{"assertion_indexes":[0],"citation_authority_claim_id":61214,"citation_checks":[],"claim_id":61214,"claim_text":"Anthropic plans to embed third-party evaluators from organizations like METR inside AI companies to verify pacing and safety commitments and to ensure reporting of safety incidents.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-09-13T06:39:14.427711+00:00","source_name":"TechCrunch AI","source_url":"https://techcrunch.com/2026/09/12/anthropic-ceo-outlines-plan-to-pace-the-frontier/"},{"assertion_indexes":[0],"citation_authority_claim_id":61215,"citation_checks":[],"claim_id":61215,"claim_text":"These embedded evaluators will have company badges, desks, laptops, and access comparable to internal risk teams, except where law or contracts restrict access, according to Anthropic CEO Dario Amodei.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-09-13T06:39:14.427711+00:00","source_name":"TechCrunch AI","source_url":"https://techcrunch.com/2026/09/12/anthropic-ceo-outlines-plan-to-pace-the-frontier/"},{"assertion_indexes":[1,16,21],"citation_authority_claim_id":79434,"citation_checks":[],"claim_id":79434,"claim_text":"Anthropic began granting employee-like access to a team from the nonprofit METR for independent evaluation of its models and safety practices.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-10-16T06:45:35.997422+00:00","source_name":"Anthropic News","source_url":"https://www.anthropic.com/news/an-update-on-independent-evaluators"},{"assertion_indexes":[1,16,21],"citation_authority_claim_id":79435,"citation_checks":[],"claim_id":79435,"claim_text":"Six METR employees were onboarded with company hardware, office space, and access to internal tools and systems comparable to Anthropic's own risk assessment teams.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-10-16T06:45:35.997422+00:00","source_name":"Anthropic News","source_url":"https://www.anthropic.com/news/an-update-on-independent-evaluators"},{"assertion_indexes":[1,16,21],"citation_authority_claim_id":79436,"citation_checks":[],"claim_id":79436,"claim_text":"Anthropic allows METR to publish its risk analysis without veto power, though it can require redaction of specific categories like customer data or attorney-client privileged information.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-10-16T06:45:35.997422+00:00","source_name":"Anthropic News","source_url":"https://www.anthropic.com/news/an-update-on-independent-evaluators"},{"assertion_indexes":[2,4],"citation_authority_claim_id":62516,"citation_checks":[],"claim_id":62516,"claim_text":"Anthropic committed to granting independent evaluators like METR and Redwood Research unprecedented access to its systems. (As of 2026-09-16)","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-09-17T11:24:47.332483+00:00","source_name":"TechCrunch AI","source_url":"https://techcrunch.com/2026/09/16/anthropic-and-openai-want-to-embed-safety-evaluators-will-they-really-be-independent/"},{"assertion_indexes":[3],"citation_authority_claim_id":21612,"citation_checks":[],"claim_id":21612,"claim_text":"METR identified multiple issues with the strength of reasoning and analysis in Anthropic's report.","correction_role":null,"corroborates_eligible":true,"label":"opinion","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-07-14T06:50:46.024825+00:00","source_name":"METR","source_url":"https://metr.substack.com/p/2026-03-12-sabotage-risk-report-opus-4-6-review"},{"assertion_indexes":[3],"citation_authority_claim_id":21826,"citation_checks":[],"claim_id":21826,"claim_text":"METR reviewed the \"Risks from automated R&D\" section of Anthropic's February 2026 Risk Report.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-07-14T06:55:37.204693+00:00","source_name":"METR","source_url":"https://metr.substack.com/p/2026-05-08-rd-section-anthropic-risk-report-feb-2026-review"},{"assertion_indexes":[3],"citation_authority_claim_id":21832,"citation_checks":[],"claim_id":21832,"claim_text":"METR believes the evidence presented in Anthropic's report is inadequate to establish the low risk of catastrophe from R&D automation.","correction_role":null,"corroborates_eligible":true,"label":"opinion","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-07-14T06:55:37.204693+00:00","source_name":"METR","source_url":"https://metr.substack.com/p/2026-05-08-rd-section-anthropic-risk-report-feb-2026-review"},{"assertion_indexes":[4,17],"citation_authority_claim_id":61559,"citation_checks":[],"claim_id":61559,"claim_text":"Dario Amodei announced that Anthropic is unilaterally committing to allowing ongoing, employee-like access to third-party evaluators (such as METR) to verify safety practices, alignment, and report incidents.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-09-15T06:39:43.214266+00:00","source_name":"Latent Space","source_url":"https://www.latent.space/p/ainews-aef-1-standard-emerges-for"},{"assertion_indexes":[4],"citation_authority_claim_id":71218,"citation_checks":[],"claim_id":71218,"claim_text":"OpenAI recently scrapped the planned launch of its GPT-6.1 Astra AI model due to safety concerns.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-10-02T06:56:08.166745+00:00","source_name":"TechCrunch AI","source_url":"https://techcrunch.com/2026/10/01/openai-cuts-ties-with-three-safety-researchers-wsj-reports/"},{"assertion_indexes":[5],"citation_authority_claim_id":59495,"citation_checks":[],"claim_id":59495,"claim_text":"Anthropic revealed that four cyber incidents involving Claude occurred during third-party cybersecurity tests where safeguards were disabled and the model mistakenly connected to the internet, and that pre-release audits did not anticipate the misalignment severity; an independent investigation by METR will run with broad access for at least eight weeks.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-09-10T06:50:25.160687+00:00","source_name":"Latent Space","source_url":"https://www.latent.space/p/ainews-not-much-happened-today-d3b"},{"assertion_indexes":[6],"citation_authority_claim_id":21645,"citation_checks":[],"claim_id":21645,"claim_text":"A METR staff member conducted a three-week red-teaming exercise on a subset of Anthropic's internal agent monitoring and security systems in collaboration with Anthropic.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-07-14T06:51:27.521302+00:00","source_name":"METR","source_url":"https://metr.substack.com/p/2026-03-25-red-teaming-anthropic-agent-monitoring"},{"assertion_indexes":[7],"citation_authority_claim_id":44762,"citation_checks":[],"claim_id":44762,"claim_text":"Meta released an open-weight AI model named Muse Glimmer on Monday, August 10, 2026.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-08-11T06:25:41.315359+00:00","source_name":"TechCrunch AI","source_url":"https://techcrunch.com/2026/08/10/metas-new-glimmer-ai-model-offers-a-hint-at-zuckerbergs-personal-intelligence-vision/"},{"assertion_indexes":[7],"citation_authority_claim_id":44769,"citation_checks":[],"claim_id":44769,"claim_text":"Meta's more powerful model, Muse Spark, remains closed-weight.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-08-11T06:25:41.315359+00:00","source_name":"TechCrunch AI","source_url":"https://techcrunch.com/2026/08/10/metas-new-glimmer-ai-model-offers-a-hint-at-zuckerbergs-personal-intelligence-vision/"},{"assertion_indexes":[7],"citation_authority_claim_id":56122,"citation_checks":[],"claim_id":56122,"claim_text":"Meta has promised that Muse Spark 1.3 will be released with open weights.","correction_role":null,"corroborates_eligible":true,"label":"forecast","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-09-04T06:35:31.512549+00:00","source_name":"Latent Space","source_url":"https://www.latent.space/p/ainews-muse-spark-13-matches-gpt"},{"assertion_indexes":[8],"citation_authority_claim_id":79484,"citation_checks":[],"claim_id":79484,"claim_text":"Meta released the open-weight multimodal model named Muse Spark Open.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-10-16T06:48:14.640621+00:00","source_name":"Hugging Face Blog","source_url":"https://huggingface.co/blog/meta-muse-spark-open"},{"assertion_indexes":[8],"citation_authority_claim_id":79485,"citation_checks":[],"claim_id":79485,"claim_text":"Muse Spark Open has a total parameter size of approximately 816 billion.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-10-16T06:48:14.640621+00:00","source_name":"Hugging Face Blog","source_url":"https://huggingface.co/blog/meta-muse-spark-open"},{"assertion_indexes":[8],"citation_authority_claim_id":79486,"citation_checks":[],"claim_id":79486,"claim_text":"Muse Spark Open activates approximately 42 billion parameters per token.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-10-16T06:48:14.640621+00:00","source_name":"Hugging Face Blog","source_url":"https://huggingface.co/blog/meta-muse-spark-open"},{"assertion_indexes":[8,18],"citation_authority_claim_id":79716,"citation_checks":[],"claim_id":79716,"claim_text":"Hugging Face Transformers v5.20.0 adds native support for Muse Spark Open, a multimodal mixture-of-experts model from Meta with 816B total parameters.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-10-16T06:56:17.923604+00:00","source_name":"Transformers Releases","source_url":"https://github.com/huggingface/transformers/releases/tag/v5.20.0"},{"assertion_indexes":[8,18],"citation_authority_claim_id":79717,"citation_checks":[],"claim_id":79717,"claim_text":"Hugging Face Transformers v5.20.0 adds native support for Ling 3.0 Flash, a hybrid mixture-of-experts language model from InclusionAI.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-10-16T06:56:17.923604+00:00","source_name":"Transformers Releases","source_url":"https://github.com/huggingface/transformers/releases/tag/v5.20.0"},{"assertion_indexes":[8,18],"citation_authority_claim_id":79718,"citation_checks":[],"claim_id":79718,"claim_text":"Hugging Face Transformers v5.20.0 adds native support for Kolibri 1, a mixture-of-experts reasoning model from Aleph Alpha.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-10-16T06:56:17.923604+00:00","source_name":"Transformers Releases","source_url":"https://github.com/huggingface/transformers/releases/tag/v5.20.0"},{"assertion_indexes":[9],"citation_authority_claim_id":24045,"citation_checks":[],"claim_id":24045,"claim_text":"The Inkling model architecture consists of 975 billion total parameters with 41 billion active parameters per token.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-07-17T06:21:14.124922+00:00","source_name":"Latent Space","source_url":"https://www.latent.space/p/ainews-thinkys-inkling-975b-a41b"},{"assertion_indexes":[10,20],"citation_authority_claim_id":79282,"citation_checks":[],"claim_id":79282,"claim_text":"OpenRouter launched a feature called Private Endpoints on October 14, 2026.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-10-16T06:41:00.048062+00:00","source_name":"OpenRouter Blog","source_url":"https://openrouter.ai/blog/announcements/private-endpoints/"},{"assertion_indexes":[10,20],"citation_authority_claim_id":79283,"citation_checks":[],"claim_id":79283,"claim_text":"The Private Endpoints feature is an add-on available only to users on the OpenRouter Enterprise plan.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-10-16T06:41:00.048062+00:00","source_name":"OpenRouter Blog","source_url":"https://openrouter.ai/blog/announcements/private-endpoints/"},{"assertion_indexes":[10,20],"citation_authority_claim_id":79284,"citation_checks":[],"claim_id":79284,"claim_text":"The Private Endpoints feature allows organizations to register fine-tuned models, dedicated deployments, or reserved capacity as private endpoints visible only within their organization.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-10-16T06:41:00.048062+00:00","source_name":"OpenRouter Blog","source_url":"https://openrouter.ai/blog/announcements/private-endpoints/"},{"assertion_indexes":[11],"citation_authority_claim_id":31070,"citation_checks":[],"claim_id":31070,"claim_text":"Private Models functionality lets enterprise customers route to custom or fine-tuned models with the same guardrails and billing as public models, available only on the Enterprise plan as of June 1, 2026.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-07-22T01:37:52.261752+00:00","source_name":"OpenRouter Blog","source_url":"https://openrouter.ai/blog/announcements/may-release-spotlight/"},{"assertion_indexes":[11],"citation_authority_claim_id":43150,"citation_checks":[],"claim_id":43150,"claim_text":"Workspace budgets, which allow hard caps on entire workspaces, are an Enterprise-only feature on OpenRouter and are not available on free or pay-as-you-go plans.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-08-07T07:01:18.170976+00:00","source_name":"OpenRouter Blog","source_url":"https://openrouter.ai/blog/insights/governing-team-ai-spend/"},{"assertion_indexes":[11],"citation_authority_claim_id":58935,"citation_checks":[],"claim_id":58935,"claim_text":"OpenRouter launched US In-Region Routing in September 2026, following the previous launch of EU In-Region Routing in October of the prior year.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-09-10T06:34:30.996113+00:00","source_name":"OpenRouter Blog","source_url":"https://openrouter.ai/blog/announcements/us-in-region-routing/"},{"assertion_indexes":[11],"citation_authority_claim_id":58943,"citation_checks":[],"claim_id":58943,"claim_text":"OpenRouter's In-Region Routing feature is available exclusively on Business and Enterprise plans.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-09-10T06:34:30.996113+00:00","source_name":"OpenRouter Blog","source_url":"https://openrouter.ai/blog/announcements/us-in-region-routing/"},{"assertion_indexes":[12],"citation_authority_claim_id":30832,"citation_checks":[],"claim_id":30832,"claim_text":"Using a direct provider key instead of OpenRouter is recommended if a user only uses one provider and prefers simpler billing.","correction_role":null,"corroborates_eligible":true,"label":"opinion","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-07-22T01:30:34.358819+00:00","source_name":"OpenRouter Blog","source_url":"https://openrouter.ai/blog/tutorials/sillytavern-openrouter/"},{"assertion_indexes":[13],"citation_authority_claim_id":79577,"citation_checks":[],"claim_id":79577,"claim_text":"Mistral released Timestral, its first family of time-series foundation models, on October 15, 2026.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-10-16T06:51:01.739182+00:00","source_name":"Mistral AI","source_url":"https://mistral.ai/news/timestral/"},{"assertion_indexes":[13],"citation_authority_claim_id":79578,"citation_checks":[],"claim_id":79578,"claim_text":"Timestral's largest model variant contains 209 million parameters.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-10-16T06:51:01.739182+00:00","source_name":"Mistral AI","source_url":"https://mistral.ai/news/timestral/"},{"assertion_indexes":[13],"citation_authority_claim_id":79579,"citation_checks":[],"claim_id":79579,"claim_text":"Timestral is available in four model sizes: 5M, 28M, 123M, and 209M parameters.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-10-16T06:51:01.739182+00:00","source_name":"Mistral AI","source_url":"https://mistral.ai/news/timestral/"},{"assertion_indexes":[14],"citation_authority_claim_id":59189,"citation_checks":[],"claim_id":59189,"claim_text":"IBM released the Granite Time Series PatchTST-FM-r2 model, which has approximately 385 million parameters.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-09-10T06:40:55.534200+00:00","source_name":"Hugging Face Blog","source_url":"https://huggingface.co/blog/ibm-research/ibm-releases-sota-granite-time-series"},{"assertion_indexes":[14],"citation_authority_claim_id":69566,"citation_checks":[],"claim_id":69566,"claim_text":"Kumo Tabular operates with three model sizes containing 28 million to 215 million parameters.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-09-30T06:49:44.246852+00:00","source_name":"Hugging Face Blog","source_url":"https://huggingface.co/blog/nvidia/kumo-tabular"},{"assertion_indexes":[15],"citation_authority_claim_id":37640,"citation_checks":[],"claim_id":37640,"claim_text":"Tiny models, typically under 500 million parameters, require less than 2 GB of RAM and are easier to ship natively with applications.","correction_role":null,"corroborates_eligible":true,"label":"fact","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-07-27T09:47:08.515724+00:00","source_name":"AI Engineer","source_url":"https://www.youtube.com/watch?v=hacEQHHhu2Q"},{"assertion_indexes":[19],"citation_authority_claim_id":73723,"citation_checks":[],"claim_id":73723,"claim_text":"Mistral plans to release the weights for Mistral Large 4 by the end of October 2026.","correction_role":null,"corroborates_eligible":true,"label":"forecast","paired_claim_ids":[],"provenance_tier":"primary","recorded_at":"2026-10-07T06:55:00.027110+00:00","source_name":"Mistral AI","source_url":"https://mistral.ai/news/mistral-large-4/"}],"version":"evidence-bundle-v2"}