米伦迪尔拥有发布 Fern，这是它的第一个AI模型。 米伦迪尔·费尔恩(Mirendil Fern)是一个来自scratch的编码模型，其上下文窗口可高达10M令牌。 米伦迪尔解释了这相当于大约23 MB代码，或大约50万行。 这是一个关于威震天LM规模的密码库，紧接着上下文进行。 Fern 0.1是第一个版本，由Mirendil-Base在短回合的训练后建造。 对于开发者来说，它类似，你同时装入一个完整的repo，一个完整的训练日志或agent痕迹的天数。
TL；数据
- 大小：0.1至0.5T参数(精确计数和未披露的活动参数)；上下文最多为10M令牌。
- 运行：仅托管早期访问；没有开放权重或定量构建宣布。
- 性能：强长文本检索，中包代理编码对开量模型，远远落后于前沿模型。
- 最佳：在 RAULER 单需求检索中100%，每长从 32K 至 10M 令牌(基准模式).
- 最糟糕的是：在Mirendil自己的桌子上，4.0号终点站占6.1%,Claude Opus 5.5为66.4%。
米伦迪尔·费尔恩是什么
Fern是来自AAI研发实验室Mirendil的语言模型。
Fern是针对AI的研发工作，而不是一般的聊天。 Mirendil列出了6种目标用途：
- 跨大型密码库工作
- 代理编码，运行命令直到测试通过
- 研究代码：培训循环、数据管道和评价
- 调试训练从长日志和堆积痕迹中进行
- 读取完整的agent或训练痕迹结束
- 对重复、沾染和垃圾进行预先检查
费恩怎么样？
费尔恩坐落在米伦迪尔-巴塞(Mirendil-Base)上，米伦迪尔·称一号研究员在3.5个月的时间里对其进行了预训。 计算总额低于10M美元，包括实验、agent代币支出和数据处理。 AI agents写了超过99%的代码。 研究者选择了什么尝试和何时干预。
培训数据仅来自公共来源：GitHub，公共代码，网页和公共数据集。 Mirendil 称没有使用蒸馏，没有合成数据，也没有许可证数据。 代码是从零开始写在PyTorch和JAX上。
Fern是3个计划模型中最小的。 Mirendil-Base-Large(0.5至2T参数)通过预训超过80%. Mirendil-Base-XL(2至10T)通过20%以上。
较长的背景是否真的有用？
米伦迪尔在自己的储存库中测试了下个托肯的预测，被压缩成42.9 MB流。 它以每字节的位数来测量交叉纵横，在位数较低的地方比较好。 这里的窗口大小是用字节而不是符号来测量的。
Fern的损失从32K字节窗口的0.246下降到16M和23M字节的0.155. 其他型号的得分高达约1M个令牌，其广告限制。 在1M字节窗口上，Kimi-K3得分0.140和Fern0.165. 所以Fern不是同等上下文中最好的预测者。 它的优点是曲线一直经过别人停车的地方。
费恩对卢拉的得分如何？
RULAR是一个标准的长文基准测试。 米伦迪尔在前期训练模式米伦迪尔-巴塞(英语：Mirendil-Base)上执行5项RULAR任务。 它使用贪婪的解码，思考和精确的字符串匹配。
寻觅坚固。 多跳推理在10M时下降到45%. 每个10M细胞只使用20个实例，而其他地方则使用100个实例。 Mirendil 称更广泛的长文成果即将到来。
Fern 0.1对代理基准测试怎么样？
米伦迪尔将Fern 0.1与开放式和前沿型号进行比较。 其他型号的得分来自型号牌，领头牌或Mirendil自己的跑步。
Fern 0.1在5个测试中的4个测试中击败了Inkling和Nemotron 3 Ultra. 它将Beam连接在SWE-Pro v2 Hard上76.5. 它在SWE-Bench Verife上追踪到Beam,73.4对81.6. 桌子上的每一个前沿和大型的开放模型都会引领它。
费尔恩如何与类似的模型进行比较？
* 米伦迪尔的公告图图图。
关键外卖
- Fern 0.1支持最多10M令牌，约23MB代码。
- 23M字节窗口每字节的比特数持续下降到0.155
- RULAR单针保持从32K到10M的百分率。
- 代理编码滞后：6.1%在终端座椅4.0.
- 由1名研究人员建造的基座模型加上1 000美元以下的agents。
请检查fern公告，Mirendil-Base预训站，并要求提前访问。 并加入我们的通讯。 等等你在电报里吗现在你也可以加入我们的电报。
需要和我们合作推广您的 GitHub Repo OR Hugging Face Page OR Product Release OR Webinar 等吗？联系我们
米哈尔·萨特(Michal Sutter)是一名数据科学专业人员，拥有帕多瓦大学的数据科学硕士学位。 在统计分析、机器学习和数据工程方面有一个坚实的基础，Michal在将复杂的数据集转化为可操作的见解方面表现突出。
本文由 AI图鉴自动抓取并翻译，版权归原作者所有；如需查看原文请点下方链接。
Mirendil has released Fern, its first AI model. Mirendil Fern is a from-scratch coding model with a context window of up to 10M tokens. Mirendil explains how that equals about 23 MB of code, or roughly half a million lines. That is a codebase on the scale of Megatron-LM, held in context at once. Fern 0.1 is the first version, built from Mirendil-Base after a short round of post-training. For developers, its like, you load a whole repo, a full training log or days of agent traces at once.
TL;DR
- Size: 0.1 to 0.5T parameters (exact count and active params not disclosed); context up to 10M tokens.
- Runs on: Hosted early access only; no open weights or quantized builds announced.
- Performance: Strong long-context retrieval, mid-pack agentic coding versus open-weight models, well behind frontier models.
- Best: 100% on RULER single-needle retrieval at every length from 32K to 10M tokens (base model).
- Worst: 6.1% on Terminal Bench 4.0, versus 66.4% for Claude Opus 5.5 in Mirendil’s own table.
What is Mirendil Fern?
Fern is a language model from Mirendil, an AI R&D lab.
Fern is aimed at AI R&D work, not general chat. Mirendil lists 6 target uses:
- Working across larger codebases
- Agentic coding that runs commands until tests pass
- Research code: training loops, data pipelines and evaluations
- Debugging training runs from long logs and stack traces
- Reading full agent or training traces end to end
- Auditing pretraining shards for duplicates, contamination and junk
How was Fern built?
Fern sits on Mirendil-Base, which Mirendil says one researcher pretrained in 3.5 months. Total compute was under $10M, including experiments, agent token spend and data processing. AI agents wrote over 99% of the code. The researcher chose what to try and when to intervene.
Training data came only from public sources: GitHub, public code, web pages and public datasets. Mirendil says it used no distillation, no synthetic data and no licensed data. The code was written from scratch on PyTorch and JAX.
Fern is the smallest of 3 planned models. Mirendil-Base-Large (0.5 to 2T parameters) is over 80% through pretraining. Mirendil-Base-XL (2 to 10T) is over 20% through.
Does longer context actually help?
Mirendil tested next-token prediction on its own repositories, concatenated into a 42.9 MB stream. It measured cross-entropy in bits per byte, where lower is better. Window sizes here are measured in bytes, not tokens.
Fern’s loss fell from 0.246 at a 32K-byte window to 0.155 at 16M and 23M bytes. The other models were scored up to about 1M tokens, their advertised limit. At a 1M-byte window, Kimi-K3 scored 0.140 and Fern 0.165. So Fern is not the best predictor at equal context. Its advantage is that the curve keeps going past where others stop.
How does Fern score on RULER?
RULER is a standard long-context benchmark. Mirendil ran 5 RULER tasks on Mirendil-Base, the pre-post-training model. It used greedy decoding, thinking off and exact string matching.
Retrieval holds up well. Multi-hop reasoning drops to 45% at 10M. Each 10M cell uses only 20 examples, versus 100 elsewhere. Mirendil says broader long-context results are coming.
How does Fern 0.1 do on agentic benchmarks?
Mirendil compared Fern 0.1 with open-weight and frontier models. Scores for other models come from model cards, leaderboards or Mirendil’s own runs.
Fern 0.1 beats Inkling and Nemotron 3 Ultra on 4 of 5 tests. It ties Beam on SWE-Pro v2 Hard at 76.5. It trails Beam on SWE-Bench Verified, 73.4 versus 81.6. Every frontier and large open model in the table leads it.
How does Fern compare with similar models?
*Figures from Mirendil’s announcement charts.
Key Takeaways
- Fern 0.1 supports up to 10M tokens, about 23 MB of code.
- Bits per byte keeps falling to 0.155 at a 23M-byte window.
- RULER single needle stays at 100% from 32K to 10M tokens.
- Agentic coding lags: 6.1% on Terminal Bench 4.0.
- Base model built by 1 researcher plus agents for under $10M.
Check out the Fern announcement, the Mirendil-Base pretraining post, and request early access. Also, feel free to follow us on Twitter and don’t forget to join our 150k+ML SubReddit and Subscribe to our Newsletter. Wait! are you on telegram? now you can join us on telegram as well.
Need to partner with us for promoting your GitHub Repo OR Hugging Face Page OR Product Release OR Webinar etc.? Connect with us
Michal Sutter is a data science professional with a Master of Science in Data Science from the University of Padova. With a solid foundation in statistical analysis, machine learning, and data engineering, Michal excels at transforming complex datasets into actionable insights.
Auto-fetched and translated by AI Atlas. All rights belong to the original author.