This frozen page shows Augur's claims and source links for one sent dispatch. Stored spot-checks appear only where the frozen edition supports them; absence is not presented as verification.
Back then it pitched its 2.6-billion-parameter model as competitive with models four times its size on tool use and multi-step agent tasks, and its 3-billion vision model as level with two rival small models, Gemma-4-E2B and Qwen3.5-2B, on tool-use tests Hugging Face Blog Hugging Face Blog.
Assertion status: No spot-check verdict is published for this assertion.
Liquid AI claims LFM2.5-2.6B is competitive with models four times its size on tool use, instruction following, and multi-step agentic tasks.
Claim 42145Label: opinionProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
Assertion 2
Liquid concedes that larger models still hold the edge on knowledge-heavy tests such as GPQA Diamond, HLE, and AA-Omniscience, which measure what a model knows rather than what it can do Hugging Face Blog.
Assertion status: No spot-check verdict is published for this assertion.
Liquid AI's LFM2.5-24B-A2B model scores 48.09 on the LiveClaw benchmark, outperforming Qwen3.6-35B-A3B's score of 38.19.
Claim 77999Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
Liquid AI reports that LFM2.5-24B-A2B leads or matches competitors on five of eight tool-use benchmarks, specifically AutomationBench, BFCLv4, ToolSandbox, LiveClaw, and τ³-Bench Banking.
Claim 78002Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
Liquid AI asserts that larger models maintain an advantage over LFM2.5-24B-A2B on knowledge-heavy benchmarks such as GPQA Diamond, HLE, and AA-Omniscience.
Claim 78003Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
Assertion 3
Liquid reports 184 tokens (the word fragments a model reads and produces) per second on an Apple M5 Max device and up to 21,000 per second on a single Nvidia B300 server chip, and in August its 2.6B model managed 220 per second on the same class of M5 Max chip while the 3B vision model hit 228 Hugging Face Blog Hugging Face Blog Hugging Face Blog.
Assertion status: No spot-check verdict is published for this assertion.
Liquid AI claims LFM2.5-2.6B is competitive with models four times its size on tool use, instruction following, and multi-step agentic tasks.
Claim 42145Label: opinionProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
Liquid AI reports that LFM2.5-24B-A2B leads or matches competitors on five of eight tool-use benchmarks, specifically AutomationBench, BFCLv4, ToolSandbox, LiveClaw, and τ³-Bench Banking.
Claim 78002Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
Liquid AI asserts that larger models maintain an advantage over LFM2.5-24B-A2B on knowledge-heavy benchmarks such as GPQA Diamond, HLE, and AA-Omniscience.
Claim 78003Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
Assertion 4
Ollama's latest release candidate (a test build ahead of a stable version) ports Apple-specific kernels, the low-level code that runs the math on a chip, to CUDA, Nvidia's chip software, and llama.cpp, another free local-model runner, just fixed a crash where machines reporting zero free memory produced invalid layer splits and aborted loading Ollama Releases llama.cpp Releases.
Assertion status: No spot-check verdict is published for this assertion.
Prior to the b11626 release, when all devices reported zero free memory, split points were normalized by a zero sum, causing NaN splits that sent layers to invalid device indices and aborted loading with a vector::_M_range_check error.
Claim 78116Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
The b11626 release of llama.cpp implements a fix to split layers evenly when no device reports free memory, addressing the previous zero-sum normalization issue.
Claim 78117Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
Assertion 5
The firm pointed 25 package managers (the tools that pull code libraries into a project) at a compromised repository, and most of them would not catch a swapped file by default, because fewer than half ship with a lockfile (the record of the exact versions a project expects) that checks for tampering, and only two of the 25 checked every single download they pulled Trail of Bits.
Assertion status: No spot-check verdict is published for this assertion.
The author tested 25 popular package managers for their ability to detect compromised repositories injecting malicious code.
Claim 78170Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
Assertion 6
A self-spreading worm that began with one hijacked maintainer account reached more than 400 packages on npm, the main public store of JavaScript code libraries Microsoft Security Blog Wiz Research.
Assertion status: No spot-check verdict is published for this assertion.
The attack resulted from the compromise of a GitHub maintainer account, which led to the publication of malicious package versions.
Claim 42082Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
Assertion 7
On access, Google is widening experimental use of its Geospatial Reasoning agent through Ask Google Earth and an Agent-to-Agent service, a protocol that lets your own agent hand a question to Google's, and the agent lets Gemini query Google's weather, flood, population, and satellite models alongside its mapping and public datasets Google Research Blog.
Assertion status: No spot-check verdict is published for this assertion.
Google Research is expanding experimental access to its Geospatial Reasoning agent via Ask Google Earth and an Agent-to-Agent (A2A) service.
Claim 78345Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
The Geospatial Reasoning agent connects Gemini to Earth AI foundation models including WeatherNext, Flood Forecasting, the Population Dynamics Foundation Model (PDFM), AlphaEarth Foundations, and Remote Sensing Foundations.
Claim 78346Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
Assertion 8
On cost, Ai2's HiRO-ACEv2 simulates rare rain, wind, and pressure extremes at 3 km resolution for about 1/300th the cost of the X-SHiELD atmospheric model Allen Institute for AI.
Assertion status: No spot-check verdict is published for this assertion.
Allen AI introduced HiRO-ACEv2, an update to its High Resolution Output Ai2 Climate Emulator system.
Claim 78134Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
Assertion 9
Claude already handles that narrow slice with a specific behavior, a banner that sends voters to TurboVote, and that is the useful part for anyone outside Anthropic: the company has a stated, election-specific response ready for the 0.51 percent Anthropic News Anthropic News Princeton CITP - Freedom to Tinker.
Assertion status: No spot-check verdict is published for this assertion.
Anthropic updated its Usage Policy to clarify rules regarding Claude's long-term, independent work capabilities and new misuse patterns in influence operations, weapons development, and surveillance.
Claim 75170Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
Anthropic updated its surveillance policy to explicitly prohibit tracking people without consent and using Claude to decide or recommend law enforcement actions like arrests.
Claim 75174Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
Anthropic’s Claude provides an election banner directing users to TurboVote and strives for political even-handedness by avoiding unsolicited political opinions.
Claim 75286Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
Anthropic analyzed 250,000 Claude conversations over a one-week period using Anthropic Insights to understand how people use the model for elections and politics. (fact, 2026-10-14T00:00:00+00:00, null, null)
Claim 78038Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
Approximately 5.75% of all Claude conversations in the United States touch broadly on politics, public policy, or social issues. (fact, 2026-10-14T00:00:00+00:00, null, null)
Claim 78039Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
Assertion 10
- Liquid AI's five-of-eight lead-or-tie on tool-use tests is self-reported; independent reruns of LiveClaw and BFCLv4 on the 24B-A2B model will show whether the on-device number travels. Hugging Face Blog
Assertion status: No spot-check verdict is published for this assertion.
Liquid AI's LFM2.5-24B-A2B model scores 48.09 on the LiveClaw benchmark, outperforming Qwen3.6-35B-A3B's score of 38.19.
Claim 77999Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
Liquid AI reports that LFM2.5-24B-A2B leads or matches competitors on five of eight tool-use benchmarks, specifically AutomationBench, BFCLv4, ToolSandbox, LiveClaw, and τ³-Bench Banking.
Claim 78002Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
Liquid AI asserts that larger models maintain an advantage over LFM2.5-24B-A2B on knowledge-heavy benchmarks such as GPQA Diamond, HLE, and AA-Omniscience.
Claim 78003Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
Assertion 11
- Ollama's CUDA port of its Apple-chip kernels is still a release candidate; a stable 0.40.5 would bring the ported Apple kernels to Nvidia machines. Ollama Releases
Assertion status: No spot-check verdict is published for this assertion.
Ollama version v0.40.5-rc0 includes a port of MLX Metal custom kernels to CUDA by @dhiltgen.
Claim 78445Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
Assertion 12
- Trail of Bits found only two package managers that verify every download; the next move is which ecosystems turn integrity checks on by default. Trail of Bits
Assertion status: No spot-check verdict is published for this assertion.
The author tested 25 popular package managers for their ability to detect compromised repositories injecting malicious code.
Claim 78170Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
Assertion 13
- Google calls Geospatial Reasoning access experimental; the Augur call above tracks whether that becomes general availability by March. Google Research Blog
Assertion status: No spot-check verdict is published for this assertion.
Google Research is expanding experimental access to its Geospatial Reasoning agent via Ask Google Earth and an Agent-to-Agent (A2A) service.
Claim 78345Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
The Geospatial Reasoning agent connects Gemini to Earth AI foundation models including WeatherNext, Flood Forecasting, the Population Dynamics Foundation Model (PDFM), AlphaEarth Foundations, and Remote Sensing Foundations.
Claim 78346Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
Assertion 14
- Anthropic's 0.51 percent election share is a mid-October baseline; a follow-up count closer to the vote would show whether queries climb. Anthropic News
Assertion status: No spot-check verdict is published for this assertion.
Anthropic analyzed 250,000 Claude conversations over a one-week period using Anthropic Insights to understand how people use the model for elections and politics. (fact, 2026-10-14T00:00:00+00:00, null, null)
Claim 78038Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
Approximately 5.75% of all Claude conversations in the United States touch broadly on politics, public policy, or social issues. (fact, 2026-10-14T00:00:00+00:00, null, null)
Claim 78039Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
Assertion 15
Every benchmark and speed figure for LFM2.5-24B-A2B comes from Liquid AI's own post and is unaudited in the evidence, and the LiveClaw comparison is against a single named rival Hugging Face Blog.
Assertion status: No spot-check verdict is published for this assertion.
Liquid AI's LFM2.5-24B-A2B model scores 48.09 on the LiveClaw benchmark, outperforming Qwen3.6-35B-A3B's score of 38.19.
Claim 77999Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
Liquid AI reports that LFM2.5-24B-A2B leads or matches competitors on five of eight tool-use benchmarks, specifically AutomationBench, BFCLv4, ToolSandbox, LiveClaw, and τ³-Bench Banking.
Claim 78002Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded
No stored spot-check names this claim in this edition.
Liquid AI asserts that larger models maintain an advantage over LFM2.5-24B-A2B on knowledge-heavy benchmarks such as GPQA Diamond, HLE, and AA-Omniscience.
Claim 78003Label: factProvenance: primaryEligible for corroboration · eligibility does not establish truthRecorded