webAI有发布 TwIL-8B，一个8.8B的数学推理模型，代码和教学遵循。 TwIL-8B推理模型是从IBM的Granite 4.2 8B训练出来的。 重量以非商业许可证在Hugging Face上。 WebAI 的发布贴出每个任务的成本，4位构件为4.98 GiB，大小为CPU或适中-VRAM GPU推论。
TL；数据
- 大小：8.79B参数，密集度。 131,072-托肯语境，从基模继承。
运行于：CPU 或适中 VRAM GPU 通过 4.98 GiB Q4 K M GGUF. 全bf16重量为16.38 GiB.
- 性能：在数学和代码上领导其大小班，在知识基准测试上进行跟踪。
- 最佳：MATH-500,98.2分，是WebAI比较中的最高分。
- 最差：GPQA Diamond70.2，比Qwen3.5-9B低11.5点，HLE为12.6(动词变体).
什么是TwIL -8B吗？
TwIL-8B是一个来自webAI的密集解码器唯一推理模型。 它有40层，隐藏大小4096，以及32个注意力头和8个KV头。 投入和产出只有英文本。 模型在每个答案前写一个“思考”块。
它以两个变体搭载，具有相同的建筑和装饰器。 平衡是默认，产生较短的推理痕迹。 华丽的原因更长，并针对硬性竞争风格的问题。
TwIL-8B扩展了WebAI的TwIL家族。 首批成员是2026年8月的1.7B型和3B型正式逻辑模型发布。
TwIL-8B是怎么训练的？
WebAI使用三个培训后阶段：
- 关于数学、STEM和随后的教学推理痕迹的全参数SFT。
- GRPO/DAPO加强数学、密码、STEM和教学提示方面的学习。 这会产生动词变体。
- 长视的RL 奖励，有利于缩短每个域长度参考的痕迹。 这就产生了平衡的变体。
培训池被除去污染，与AIME、HMMT、MATH-500、GPQA和其他电极相匹配13克。
TwIL-8B如何在基准测试上表现？.
所有 TwIL-8B 的分数都是自报的，高功率的通过@ 1，并启用思维功能。 在动词变体上只测量了标有Q的分数。
- 数学：TwIL-8B在MATH-500上得分98.2. Qwen3.5-9B在第三方领导板上列出96.6分。 Gemma 4 31B在Red Hat的轻压跑中打进87.9分。 在AIME 2025上，TwIL-8B为88.3对87.1的Gemma 4 31B. 在AIME 2026上，它得分88.3分，后Qwen3.5-9B分以92.5分，Gemma431B分以89.2分。
- 代码：LiveCode Bench†为81.6,Qwen3.5-9B为65.6,Gemma 4 31B为80. WebAI 指出其LiveCode Bench版本没有录制。 竞争对手得分为v6.
- 指示如下：IF Bench†在64.5时比Qwen3.5-9B高73.4,9.9分。 IFEval为一级，为91.8对91.5.
——知识：这是弱点。 GPQA钻石为70.2,Qwen3.5-9B为81.7,Gemma 4 31B为84.3。 MMLU-Pro†为76.5对82.5,Qwen3.5-9B为8.5。
横跨13个共享的基准测试，TwIL-8B对Qwen3.5-9B，对Gemma 4 31B，约3.5x其大小，走3-11.
TwIL-8B使用较少的推理符号吗？
是的，与其最接近的开放对等。 在AIME 2026上，TwIL-8B平均每个问题约19.5k输出令牌。 Qwen3.5-9B使用31.2k，因此TwIL-8B使用率减少了38%. 得分4.2分低点。 在2026年2月的HMMT，差距为32.9k对40.6k，即减少19%。
边疆模式仍然更加精简。 WebAI引用了GPT-5.5. 每AIME 2026年问题大约5k个令牌。 WebAI表示，使用和标注器不同，因此将此视为方向性。
TwIL-8B对Qwen3.5-9B对Gemma 4
TwIL-8B数字来自模型卡。 † = 动词变体。 除非标明，否则每个供应商自报所有分数。
你怎么在本地运行TwIL-8B?
从 pretract("webAI-Official/TwIL-8B")加载平衡变体。 传子文件夹="动词"为动词之一。 模式卡列出GGUF为llama.cpp所建：
- Q4 K M,4.98 GiB，两种变体都使用
- Q8 0,8.70 GiB，只有动词
- F16,16.38 GiB，只有动词
给它大量的预算。 该卡的示例设置了max new tokens=32768. 短期预算会缩短推理，得分更差。 所有报告得分均使用bf16重量，因此预计Q4 K M时会有小幅下降。 目前还没有主机推论提供者为Hugging Face上的模型服务。
关键外卖
- TwIL-8B是一种从IBM Granite 4.2 8B训练出来的8.79B密集推理模型。
- 在MATH-500和LiveCode Bench上张贴98.2和81.6†。
- 在AIME 2026上，它使用比Qwen3.5-9B少38%的代币。
- 知识薄弱：GPQA钻石小径Qwen3.5-9B11.5分。
- 重量是开放的，但只允许非商业用途。
请检查access-date=中的日期值 (帮助) Hugging Face和WebAI在X的公告上的型号卡和重量。 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
Asif Razzaq是Marktechpost AI Media Inc.的首席执行官。 作为一个有远见的企业家和工程师，阿西夫致力于利用人工智能的潜力促进社会福祉。 他最近的努力是人工智能媒体平台的发布，该平台以深入报道机器学习和深层学习新闻而著称，这些新闻在技术上既可靠又为广大观众所容易理解。 该平台每月有超过200万次观看，表明其受观众欢迎。
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。
webAI has released TwIL-8B, an 8.8B reasoning model for math, code and instruction following. The TwIL-8B reasoning model is post-trained from IBM’s Granite 4.2 8B. Weights are on Hugging Face under a non-commercial license. webAI’s launch post frames the pitch around cost per task, not raw scale. The 4-bit build is 4.98 GiB, sized for CPU or modest-VRAM GPU inference.
TL;DR
- Size: 8.79B parameters, dense. 131,072-token context, inherited from the base model.
- Runs on: CPU or modest-VRAM GPU via a 4.98 GiB Q4_K_M GGUF. Full bf16 weights are 16.38 GiB.
- Performance: Leads its size class on math and code, trails on knowledge benchmarks.
- Best: MATH-500 at 98.2, the top score in webAI’s comparison.
- Worst: GPQA Diamond at 70.2, 11.5 points behind Qwen3.5-9B. HLE is 12.6 (verbose variant).
What is TwIL-8B?
TwIL-8B is a dense decoder-only reasoning model from webAI. It has 40 layers, hidden size 4096, and 32 attention heads with 8 KV heads. Input and output are English text only. The model writes a <think> block before each answer.
It ships in 2 variants with the same architecture and tokenizer. Balanced is the default and produces shorter reasoning traces. Verbose reasons longer and targets hard contest-style problems.
TwIL-8B extends webAI’s TwIL family. The first members were 1.7B and 3B formal-logic models released in August 2026.
How was TwIL-8B trained?
webAI used 3 post-training stages:
- Full-parameter SFT on reasoning traces for math, STEM and instruction following.
- GRPO/DAPO reinforcement learning on math, code, STEM and instruction prompts. This produces the verbose variant.
- Length-aware RL with a reward that favors shorter traces against per-domain length references. This produces the balanced variant.
Training pools were decontaminated with 13-gram matching against AIME, HMMT, MATH-500, GPQA and other evals.
How does TwIL-8B perform on benchmarks?
All TwIL-8B scores are self-reported, high-effort pass@1 with thinking enabled. Scores marked † were measured on the verbose variant only.
- Math: TwIL-8B scores 98.2 on MATH-500. Qwen3.5-9B lists 96.6 on a third-party leaderboard. Gemma 4 31B scored 87.9 in Red Hat’s lighteval run. On AIME 2025, TwIL-8B posts 88.3 against 87.1 for Gemma 4 31B. On AIME 2026 it scores 88.3, behind Qwen3.5-9B at 92.5 and Gemma 4 31B at 89.2.
- Code: LiveCodeBench† is 81.6, against 65.6 for Qwen3.5-9B and 80.0 for Gemma 4 31B. webAI notes its LiveCodeBench version was not recorded. Competitor scores are v6.
- Instruction following: IFBench† is 74.4, 9.9 points above Qwen3.5-9B at 64.5. IFEval is level, at 91.8 vs 91.5.
- Knowledge: This is the weak spot. GPQA Diamond is 70.2, against 81.7 for Qwen3.5-9B and 84.3 for Gemma 4 31B. MMLU-Pro† is 76.5 vs 82.5 for Qwen3.5-9B.
Across 13 shared benchmarks, TwIL-8B goes 7-6 against Qwen3.5-9B. Against Gemma 4 31B, about 3.5x its size, it goes 3-11.
Does TwIL-8B use fewer reasoning tokens?
Yes, against its closest open peer. On AIME 2026, TwIL-8B averages about 19.5k output tokens per problem. Qwen3.5-9B uses 31.2k, so TwIL-8B uses 38% fewer. It scores 4.2 points lower. On HMMT Feb 2026 the gap is 32.9k vs 40.6k, or 19% fewer.
Frontier models remain far leaner. webAI cites about 5k tokens per AIME 2026 problem for GPT-5.5. webAI states that harnesses and tokenizers differ, so treat this as directional.
TwIL-8B vs Qwen3.5-9B vs Gemma 4
TwIL-8B figures from the model card. † = verbose variant. All scores self-reported by each vendor unless marked.
How do you run TwIL-8B locally?
from_pretrained("webAI-Official/TwIL-8B") loads the balanced variant. Pass subfolder="verbose" for the verbose one. The model card lists GGUF builds for llama.cpp:
- Q4_K_M, 4.98 GiB, for both variants
- Q8_0, 8.70 GiB, verbose only
- F16, 16.38 GiB, verbose only
Give it a large generation budget. The card’s example sets max_new_tokens=32768. A short budget truncates reasoning and scores far worse. All reported scores used bf16 weights, so expect small drops at Q4_K_M. No hosted inference provider serves the model on Hugging Face yet.
Key Takeaways
- TwIL-8B is an 8.79B dense reasoning model post-trained from IBM Granite 4.2 8B.
- It posts 98.2 on MATH-500 and 81.6† on LiveCodeBench.
- It uses 38% fewer tokens than Qwen3.5-9B on AIME 2026.
- Knowledge is weak: GPQA Diamond trails Qwen3.5-9B by 11.5 points.
- Weights are open but licensed for non-commercial use only.
Check out the model card and weights on Hugging Face and webAI’s announcement on X. 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
Asif Razzaq is the CEO of Marktechpost AI Media Inc.. As a visionary entrepreneur and engineer, Asif is committed to harnessing the potential of Artificial Intelligence for social good. His most recent endeavor is the launch of an Artificial Intelligence Media Platform, Marktechpost, which stands out for its in-depth coverage of machine learning and deep learning news that is both technically sound and easily understandable by a wide audience. The platform boasts of over 2 million monthly views, illustrating its popularity among audiences.
Auto-fetched and translated by AI Atlas. All rights belong to the original author.