跳到正文
智迹闻AuraTracer

发布于

LFM2.5-8B-A1B

公开成绩与计算依据

综合参考分根据已公开的评测资料整理。查看发布日期依据

综合参考分 10.4 · 9项基准

推理求解 24.5

AIME · 56.7

huggingface.co · 官方发布比较 · 2026

LFM2.5-8B-A1B · Official published comparison; AIME 2026; exact column LFM2.5-8B-A1B. Footnotes and all competing runs retained in source.

MATH-500 · 93.2

huggingface.co · 官方发布比较 · 500-item subset

LFM2.5-8B-A1B · Official published comparison; MATH-500; exact column LFM2.5-8B-A1B. Footnotes and all competing runs retained in source.

GPQA Diamond · 0.513131313131313 比例

Artificial Analysis · 独立评测 · Diamond

LFM2.5-8B-A1B · Current AA standalone protocol. Status verified in aa-method-current.md section gpqa-diamond

工具与任务执行 0.0

BFCL · 48.50

huggingface.co · 官方发布比较 · 4

LFM2.5-8B-A1B · Official published comparison; BFCLv4; exact column LFM2.5-8B-A1B. Footnotes and all competing runs retained in source.

编程 0.0

LiveCodeBench · 39.8

huggingface.co · 官方发布比较 · release_v6

LFM2.5-8B-A1B · Official published comparison; LiveCodeBench v6; exact column LFM2.5-8B-A1B. Footnotes and all competing runs retained in source.

语言与信息处理 16.1

IFBench · 0.556462585034014 比例

Artificial Analysis · 独立评测 · current published task suite

LFM2.5-8B-A1B · Artificial Analysis current published protocol; ifbench

Multi-IF · 71.4

huggingface.co · 官方发布比较 · current maintainer dataset

LFM2.5-8B-A1B · Official published comparison; Multi-IF; exact column LFM2.5-8B-A1B. Footnotes and all competing runs retained in source.

IFEval · 90.8

huggingface.co · 官方发布比较 · current maintainer dataset

LFM2.5-8B-A1B · Official published comparison; IFEval; exact column LFM2.5-8B-A1B. Footnotes and all competing runs retained in source.

MMLU-Pro · 63.1

huggingface.co · 官方发布比较 · current maintainer dataset

LFM2.5-8B-A1B · Official published comparison; MMLU-Pro; exact column LFM2.5-8B-A1B. Footnotes and all competing runs retained in source.

查看原始成绩与计算数据

资料来源

模型资料来源 · 目录许可