Unverified50% confidenceFactExact time
TimesFM 2.5 has about 200 million parameters, 20 transformer layers, a model width of 1280 with 16 attention heads, an input patch of 32 steps, a single output of 128 steps, and a maximum context of 16384 data points.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/17/2026
First Seen
Valid until: 1/15/2027
Sources
Related Entities
Related Claims
UnverifiedLFM2.5-24B-A2B uses a 40-layer architecture: 30 double-gated short convolution blocks + 10 GQA attention layers, 64 experts per MoE block, top-4 routing, a 131,072 token context, and a 128K vocabulary.68% similarUnverifiedThe author's generation setup is an RTX 5070 Ti GPU with 32GB of RAM, 20 sampling steps, approximately 53 seconds per iteration, resolution set to 0.6, aspect ratio 1:1, and no turbo LoRA used.66% similarUnverifiedBy configuring the update frequency to batch-refresh every 50ms or 100ms for example, the number of re-renders can be reduced by 5 to 20 times.66% similarUnverifiedA tester ran a 20.7B dense model on a dual-4090 (48 GB VRAM) workstation and achieved 27.6 tokens per second, only slightly faster than Dwarf Star, but with over 10 times fewer parameters.66% similarUnverifiedThe LFM2.5-VL-450M vision-language model has about 450 million parameters, a download size of 350-400MB, and can be used for real-time camera description and paused-frame question answering.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1057089API
curl https://kongchang.com/api/v1/knowledge/claims/1057089MCP
get_claim(id=1057089)