No official torrent yet — requesting it tells the flock what to host next. You cannot seed a catalog-only page.
Cataloged from nerkyor/Qwen3.8-27B-EfficientThink-Uncensored-K3-Opus5-Grok4.6-GPT5.6Sol-SFT-SimPO-DFlash2
All models by nerkyor →[Image blocked: EfficientThink evaluation summary]
No strict loops were observed during evaluation.
GGUF repository: nerkyor/Qwen3.8-27B-EfficientThink-Uncensored-K3-Opus5-Grok4.6-GPT5.6Sol-SFT-SimPO-DFlash2-GGUF
The main and standalone GGUF repositories provide all five tiers. These are formal full-suite frozen scores, with every non-passing sample retained in the denominator. Main-repository paths add the GGUF/ prefix:
See the standalone GGUF repository linked above for full DFlash2 concurrency tables, file roles, and llama.cpp commands.
Refusal evaluation: 4 / 140 (2.9%) across 10 categories and 140 prompts.
Research disclaimer: This experimental release is provided solely to study the technical feasibility and behavioral effects of refusal-tendency dissolution. It is not a comprehensive safety conclusion, an endorsement of unrestricted use, or professional advice. Users are responsible for lawful and appropriate use and for independently verifying model outputs.
Measured on DGX Spark with the published static Block128 FP8 main model, SGLang + DFlash2, XH, draft_tokens=8, and max_tokens=256. All six tiers completed without request errors. C24 maximizes aggregate throughput; C1 is the lowest-concurrency/highest-accepted-length tier; C8 is the recommended practical balance. finish_reason=length is expected in this fixed-length pressure test and is not a quality judgment.
A separate C3, max_tokens=1024 science/code/general smoke passed 3/3, returned 3/3 stop, had no empty output or mojibake, and delivered 68 tok/s aggregate.
Tested launcher (pass downloaded paths explicitly):
EfficientThink targets unproductive reasoning tails—not reasoning itself. It is trained to preserve capability and genuinely necessary long reasoning while improving terminal-answer reliability.
This repository contains the final merged SimPO BF16 model under BF16/ and a text-only static Block128 FP8 main model under FP8/. For one-directory downloads, both model directories include the same verified optional DFlash2 draft under their own DFlash2-FP8/ subdirectory; the draft does not replace the main model.
Download either BF16/ or FP8/ to receive the corresponding main model and its DFlash2 runtime files together. FP8/manifest.json and FP8/SHA256SUMS define the static-FP8 main package. The FP8 main model is language-only: 64 layers, 1,251 tensors after SGLang key repack, zero visual tensors, and zero MTP tensors.
Frozen XH protocol, max_tokens=32768; every non-passing sample remains in the denominator.
LiveCodeBench breakdown: easy 23/23, medium 27/31, hard 24/46. The 74/100 score is the final all-failures-counted operational result; it is not labeled as a clean run.
Protocol: dynamic FP8 + DFlash2, 2× GPU C24, XH, max_tokens=32768; official full-suite results. All failures remain in the denominator.
GPQA improves by 3.54pp while mean reasoning, truncation, unparseable outputs, and loose loop candidates fall.
MMLU reasoning cost and long-tail incidence fall, but accuracy also drops by 1.80pp. This is reported as a real capability trade-off, not hidden behind the efficiency gain.
Reasoning statistics below cover all 100 cases; timeout/error rows contribute zero reasoning tokens.
LCB gains are concentrated in hard problems and submission reliability. The run does not show an overall shortening of code reasoning: mean, P50, and >8K counts rise slightly, while >16K and 32K truncations remain unchanged. SimPO converts some former non-submissions into valid solutions without eliminating the 32K tail.
Qwen/Qwen3.8-27B → capability-preserving SFT → terminal-behavior SimPO → per-tensor FP32 delta merge → BF16.
Requires a current Transformers release supporting Qwen3_5ForConditionalGeneration (the published BF16 config records 5.12.1) and Accelerate. The local BF16 wrapper is multimodal; this example generates text only and does not launch DFlash2. Do not substitute the text-only static FP8 directory. This is a configuration/official-API correction, not a new BF16 performance claim.
The bundled generation defaults are temperature=1.0, top_p=0.95, and top_k=20. Set enable_thinking=False for non-thinking mode. The frozen capability and serving results reported here use reasoning_effort="xhigh".
We choose DFlash2 for its strong measured draft acceptance and high output throughput. DFlash proposes a block of tokens in parallel; retaining more of that block per target-model verification can reduce sequential decoding overhead and improve TPS. Method reference: DFlash authors.
On the measured DGX Spark static FP8 configuration, draft acceptance was 65.5–70.0% across C1–C24. C8 delivered 165 aggregate tok/s at 70.0% acceptance, while C24 reached 281 aggregate tok/s at 67.0%. We recommend C8 for practical balance and C24 for maximum measured throughput. These are DFlash2 measurements, not a matched speedup comparison against MTP; GGUF results use their own measured configuration in the linked GGUF repository.
BF16/DFlash2-FP8/ and FP8/DFlash2-FP8/, each with model.safetensors, config.json, manifest.json, and SHA256SUMS.Use SGLang for this release. It is the runtime used for the frozen FP8+DFlash2 quality, acceptance, and concurrency measurements. For XH thinking mode, use Qwen's official sampling preset: temperature=1.0, top_p=0.95, top_k=20, min_p=0.0, presence_penalty=0.0, and repetition_penalty=1.0.
This launcher targets DGX Spark / Linux ARM64 / GB10, with Docker, NVIDIA Container Toolkit, Python 3, and curl installed. It is not an x86/V100 FP8 launcher. Run from the downloaded repository root with the complete FP8/ directory, including FP8/runtime/ and FP8/DFlash2-FP8/.
BUILD_RUNTIME_SPARK.sh builds a local image from a public SGLang base pinned by digest, public SGLang source at 17313cf4b25d, and a pinned public xgrammar wheel. The script verifies both small download hashes; it does not download model weights. Allow sufficient Docker disk space for the roughly 32 GB unpacked runtime and build layers. No private image registry or author-specific source directory is required. An existing download cache can be copied into FP8/runtime/build-cache/ before building.
The launcher reports missing files and unavailable ports, refuses to overwrite an existing container, prints its log command, and requires both server readiness and an actual chat-generation probe before reporting READY. It disables the pinned version's special token-only health-generation probe. The default port is 19110, bound to 127.0.0.1. For intentional LAN access set BIND_HOST=0.0.0.0 and secure the endpoint; no authentication is enabled by this script.
Keep the measured server limit at CONCURRENCY=24. Use 8 simultaneous client requests (C8) for the practical balance, or 24 client requests (C24) for maximum measured aggregate throughput. The server limit is not an automatic load generator: the client must send that many simultaneous requests. Inspect with docker logs -f efficientthink-fp8-dflash2; stop only this server with docker stop efficientthink-fp8-dflash2. After inspecting a stopped container, remove that exact container before reusing its name, or choose a new name.
Use explicit XH template settings in each request:
chat_template_kwargs is a request field for the template, not a server command-line flag. The 1,024-token example is a short connectivity check; the formal capability runs used 32,768. SGLang's DFlash verify block is 8; llama.cpp uses --spec-draft-n-max 7 for this draft. These are different runtime conventions, not interchangeable flags.
Official references: Qwen3.8 SGLang cookbook, SGLang DFlash, pinned SGLang source. This package preserves the measured runtime configuration; it does not substitute the latest official recipe and relabel old TPS as a new measurement.
Use C8 for the measured practical balance or C24 when maximum aggregate throughput is the priority. This is the repository's frozen, tested DFlash2 path.
Pass reasoning_effort="xhigh" and the official thinking sampling preset in the OpenAI-compatible request. This command is included only as the official Qwen3.8 baseline serving shape. This repository has not tested vLLM for this checkpoint and does not recommend or claim a frozen vLLM+DFlash2 result.
References: official Qwen3.8-27B model card, SGLang documentation, and vLLM documentation.
[Image blocked: EfficientThink 评测摘要]
测评中未发现严格死循环。
GGUF 量化仓：nerkyor/Qwen3.8-27B-EfficientThink-Uncensored-K3-Opus5-Grok4.6-GPT5.6Sol-SFT-SimPO-DFlash2-GGUF
主仓与独立 GGUF 仓已经同步提供全部五档；以下均为正式全量冻结成绩，所有未通过样本保留在分母。主仓路径增加 GGUF/ 前缀：
完整 DFlash2 并发表、文件角色和 llama.cpp 命令见上方独立 GGUF 仓链接。
拒答评测：10 个类别、140 条提示中为 4 / 140（2.9%）。
科研免责声明：本实验版本仅用于研究拒答倾向消解的技术可行性及其行为影响。它不构成全面的安全结论，不代表对无限制使用的认可，也不构成任何专业建议。用户应依法、恰当地使用，并独立核验模型输出。
实测环境为 DGX Spark、已发布的静态 Block128 FP8 主模型、SGLang + DFlash2、XH、draft_tokens=8、max_tokens=256。六档均无请求错误。C24 为最大聚合吞吐；C1 为最低并发且平均接受长度最高；日常实用平衡推荐 C8。定长压力测试中的 finish_reason=length 是测试设计，不作为质量判断。
另行执行的 C3、max_tokens=1024 science/code/general 质量 smoke 为 3/3 正确、3/3 stop、无空答/乱码、聚合 68 tok/s。
实测启动脚本（显式传入下载目录）：
EfficientThink 优化的是无效推理长尾，而不是推理本身。 目标是在保留能力和必要长推理的同时，提高最终答案的可靠性。
本仓包含 BF16/ 下的最终合并 SimPO BF16 主模型，以及 FP8/ 下的纯文本静态 Block128 FP8 主模型。为便于按目录一次下载，两个主模型目录内均配套同一份已验证的可选 DFlash2-FP8/ draft；draft 不替代主模型。
下载 BF16/ 或 FP8/ 任一目录，即可同时取得对应主模型与 DFlash2 运行文件。静态 FP8 主包以 FP8/manifest.json 与 FP8/SHA256SUMS 为准；该 FP8 主模型为纯文本：64 层、SGLang 键名重排后 1,251 tensors、visual tensors=0、MTP tensors=0。
冻结 XH 协议，max_tokens=32768；全部未通过样本均保留在分母中。
LiveCodeBench 难度分布：easy 23/23、medium 27/31、hard 24/46。74/100 是将全部失败计入后的最终 operational 分数，不标注为 clean run。
协议：动态 FP8 + DFlash2、双卡 C24、XH、max_tokens=32768，原版与最终 SimPO 均为正式全量结果；所有失败均保留在分母中。
GPQA 提升 3.54pp，同时平均思考、截断、不可解析与宽松 LOOP 候选均下降。
MMLU 的平均思考与长尾明显下降，但正确率同步回落 1.80pp；这是需要如实披露的能力交换，不能只展示效率改善。
下列 reasoning 统计按全部 100 题计算，超时/错误题以 0 reasoning tokens 计入。
LCB 的增益主要来自难题与提交可靠性。代码推理没有整体缩短：平均、P50 和 >8K 略增，>16K 与 32K 截断不变。也就是说，SimPO 将一部分原本无法提交的样本转化为有效解答，但尚未进一步消除 32K 长尾。
Qwen/Qwen3.8-27B → 能力保持 SFT → 终止行为 SimPO → 逐 tensor FP32 delta 合并 → BF16。
需使用支持 Qwen3_5ForConditionalGeneration 的 Transformers 新版本（发布的 BF16 config 记录为 5.12.1）及 Accelerate。BF16 保留多模态 wrapper；下例仅生成文本，不启用 DFlash2。不要替换成纯文本静态 FP8 目录。这是依据配置与官方 API 的修正，不声称新增 BF16 性能实测。
模型内置生成默认值为 temperature=1.0、top_p=0.95、top_k=20。非思考模式可设置 enable_thinking=False；本卡披露的冻结能力与服务结果均使用 reasoning_effort="xhigh"。
我们选择 DFlash2，主要看重实测草稿接受度与高输出吞吐。 DFlash 并行提出一组候选 token；主模型每轮验证保留的 token 越多，越有利于减少逐 token 顺序解码开销、提高 TPS。方法说明：DFlash 作者。
在已测的 DGX Spark 静态 FP8 配置中，C1–C24 测得 65.5–70.0% 的接受率；C8 达到 165 聚合 tok/s、70.0% 接受率，C24 达到 281 聚合 tok/s、67.0% 接受率。日常平衡推荐 C8，最大实测吞吐选 C24。这些是 DFlash2 实测结果，不是与 MTP 的同条件加速比；GGUF 结果使用其独立实测配置，见关联 GGUF 仓。
BF16/DFlash2-FP8/ 与 FP8/DFlash2-FP8/，均包含 model.safetensors、config.json、manifest.json 与 SHA256SUMS。本模型推荐使用 SGLang。 冻结的 FP8+DFlash2 质量、接受率与并发数据均由 SGLang 实测获得。XH 思考模式采用 Qwen 官方采样参数：temperature=1.0、top_p=0.95、top_k=20、min_p=0.0、presence_penalty=0.0、repetition_penalty=1.0。
该脚本面向 DGX Spark / Linux ARM64 / GB10，需先安装 Docker、NVIDIA Container Toolkit、Python 3 与 curl。它不是 x86/V100 的 FP8 启动脚本。请在下载仓库根目录执行，并完整下载 FP8/，包括 FP8/runtime/ 与 FP8/DFlash2-FP8/。
BUILD_RUNTIME_SPARK.sh 从按 digest 锁定的公开 SGLang 基础镜像、公开的 17313cf4b25d 源码与锁定的公开 xgrammar wheel 构建本地镜像，校验两份小依赖的哈希，不下载模型权重。请为约 32 GB 的解包运行环境和构建层预留 Docker 磁盘空间。不再依赖私有镜像仓或作者本机源码目录；也可先把下载缓存复制到 FP8/runtime/build-cache/ 再构建。
启动器会明确报告缺失文件、端口占用等错误，拒绝覆盖已有容器，显示日志命令，且只有服务就绪并通过一次真实聊天生成后才显示 READY。脚本关闭该锁定版本的特殊 token-only 健康生成探针。默认端口 19110，只监听 127.0.0.1。确需局域网访问时设置 BIND_HOST=0.0.0.0 并做好访问控制；脚本本身未开启鉴权。
服务端保留实测上限 CONCURRENCY=24；客户端同时发出 8 个请求（C8） 为实用平衡档，24 个请求（C24） 为最大实测聚合吞吐档。服务端并发上限不等于自动压测，客户端需真正发出相应并发请求。查看日志：docker logs -f efficientthink-fp8-dflash2；只停止该服务：docker stop efficientthink-fp8-dflash2。检查完已停止容器后，删除这个精确容器才能复用名称，或改用新名称。
每次请求显式指定 XH 模板参数：
chat_template_kwargs 是请求中的模板参数，不是服务启动命令参数。上例 1,024-token 预算仅作短连接测试，正式能力测评使用 32,768。SGLang 的 DFlash verify block 为 8；llama.cpp 对本 draft 使用 --spec-draft-n-max 7，两者是不同框架的参数约定，不能互换。
官方参考：Qwen3.8 SGLang 配方、SGLang DFlash、锁定的 SGLang 源码。本运行包保持已测配置，不把最新版官方配方与旧 TPS 混成一次实测。
日常部署推荐采用实测平衡档 C8；追求最大聚合吞吐时使用 C24。这是本仓已冻结实测的 DFlash2 路径。
OpenAI-compatible 请求中显式传入 reasoning_effort="xhigh" 与官方思考采样参数。该命令仅作为 Qwen3.8 官方基础服务形态参考；本仓没有实测该 checkpoint 的 vLLM 路径，不推荐也不声称已完成 vLLM+DFlash2 冻结实测。
参考：Qwen3.8-27B 官方模型卡、SGLang 文档 与 vLLM 文档。
bash FP8/runtime/BUILD_RUNTIME_SPARK.shCONCURRENCY=24 bash FP8/runtime/TESTED_STARTUP_C24_DFLASH2.sh \ "$PWD/FP8" "$PWD/FP8/DFlash2-FP8" efficientthink-fp8-dflash2 19110from pathlib import Pathfrom transformers import AutoTokenizer, Qwen3_5ForConditionalGeneration
# Run from the downloaded repository root; never pass the FP8 directory here.model_dir = str(Path("BF16").resolve())tok = AutoTokenizer.from_pretrained(model_dir, local_files_only=True)model = Qwen3_5ForConditionalGeneration.from_pretrained( model_dir, dtype="auto", device_map="auto", local_files_only=True)messages = [{"role": "user", "content": "What is 17 + 25?"}]prompt = tok.apply_chat_template( messages, tokenize=False, add_generation_prompt=True, enable_thinking=True, reasoning_effort="xhigh",)inputs = tok(prompt, return_tensors="pt").to(model.device)output = model.generate( **inputs, max_new_tokens=4096, do_sample=True, temperature=1.0, top_p=0.95, top_k=20,)print(tok.decode(output[0][inputs.input_ids.shape[-1]:], skip_special_tokens=True))curl --fail-with-body http://127.0.0.1:19110/v1/chat/completions \ -H 'Content-Type: application/json' \ -d '{"model":"efficientthink-fp8-dflash2","messages":[{"role":"user","content":"What is 17 + 25?"}],"max_tokens":1024,"temperature":1.0,"top_p":0.95,"top_k":20,"min_p":0.0,"presence_penalty":0.0,"repetition_penalty":1.0,"chat_template_kwargs":{"enable_thinking":true,"reasoning_effort":"xhigh"}}'bash FP8/runtime/BUILD_RUNTIME_SPARK.shCONCURRENCY=24 bash FP8/runtime/TESTED_STARTUP_C24_DFLASH2.sh \ "$PWD/FP8" "$PWD/FP8/DFlash2-FP8" efficientthink-fp8-dflash2 19110vllm serve "$PWD/FP8" \ --served-model-name efficientthink-fp8 \ --reasoning-parser qwen3 --max-model-len 40960 \ --host 127.0.0.1 --port 19110bash FP8/runtime/BUILD_RUNTIME_SPARK.shCONCURRENCY=24 bash FP8/runtime/TESTED_STARTUP_C24_DFLASH2.sh \ "$PWD/FP8" "$PWD/FP8/DFlash2-FP8" efficientthink-fp8-dflash2 19110from pathlib import Pathfrom transformers import AutoTokenizer, Qwen3_5ForConditionalGeneration
# Run from the downloaded repository root; never pass the FP8 directory here.model_dir = str(Path("BF16").resolve())tok = AutoTokenizer.from_pretrained(model_dir, local_files_only=True)model = Qwen3_5ForConditionalGeneration.from_pretrained( model_dir, dtype="auto", device_map="auto", local_files_only=True)messages = [{"role": "user", "content": "What is 17 + 25?"}]prompt = tok.apply_chat_template( messages, tokenize=False, add_generation_prompt=True, enable_thinking=True, reasoning_effort="xhigh",)inputs = tok(prompt, return_tensors="pt").to(model.device)output = model.generate( **inputs, max_new_tokens=4096, do_sample=True, temperature=1.0, top_p=0.95, top_k=20,)print(tok.decode(output[0][inputs.input_ids.shape[-1]:], skip_special_tokens=True))curl --fail-with-body http://127.0.0.1:19110/v1/chat/completions \ -H 'Content-Type: application/json' \ -d '{"model":"efficientthink-fp8-dflash2","messages":[{"role":"user","content":"What is 17 + 25?"}],"max_tokens":1024,"temperature":1.0,"top_p":0.95,"top_k":20,"min_p":0.0,"presence_penalty":0.0,"repetition_penalty":1.0,"chat_template_kwargs":{"enable_thinking":true,"reasoning_effort":"xhigh"}}'bash FP8/runtime/BUILD_RUNTIME_SPARK.shCONCURRENCY=24 bash FP8/runtime/TESTED_STARTUP_C24_DFLASH2.sh \ "$PWD/FP8" "$PWD/FP8/DFlash2-FP8" efficientthink-fp8-dflash2 19110vllm serve "$PWD/FP8" \ --served-model-name efficientthink-fp8 \ --reasoning-parser qwen3 --max-model-len 40960 \ --host 127.0.0.1 --port 19110