TuTouPower

code llm

[as42 @as42tsx]

@MaxForAI 你进入 https://t.co/ee72EkEdnd 试试，笑死，OpenAI做梦都没想到已经被xAI预判了

[TuTouPower @TuTouPower]

10.1 opencode 会上架一款 red-loong 匿名模型

[TuTouPower @TuTouPower]

rust 重写一切已经不够了，接下来是汇编

[DHH @dhh]

I ported the Omarchy screensaver engine (ttfx) from Rust to x86-64 assembler, and it's up to 17x faster!! One-shot translation by Opus 5.5. We keep drilling until the agentic drill bit hits bedrock! https://t.co/SFJXcquMip

[TuTouPower @TuTouPower]

https://t.co/62JnjbLYaT 大模型测速和单价网

[TuTouPower @TuTouPower]

什么叫手机号就是人的api。打电话就可以调用我了？什么叫国家免费训练了我这个本地部署小模型训练了九年，然后再自费训练三年+四年，训练出的结果是注意力涣散加过拟合的，上下文顶多几千字，支持联网搜索，图片识别也还不错，训练结束后我这个模型出个Coding Plan，一个月四千你死命蹬，周一到周五你呼喊两三个汉字作为api，周末得输入一个11位数字作为api，随叫随到从不缺席，我要是遇上一个同样合适的本地部署小模型还能再产生一个小模型，真是太有意思了。

[TuTouPower @TuTouPower]

测试了下，grok bot 虚拟机注册 muse 失败，注册 claude 直接封号。
muse 虚拟机禁止打开 https://t.co/EUH9MlSV39 网页，注册 claude 成功。

[TuTouPower @TuTouPower]

测速代码和数据开源在 https://t.co/KHSy15epqC

[TuTouPower @TuTouPower]

做了个大模型测速的网站，https://t.co/WlqiaMlozl

mimo ultraspeed 速度一骑绝尘，不过没人用得起。
最具性价比的还是 gemini 和 deepseek 又快又便宜。

[TuTouPower @TuTouPower]

所以 luna 和 sol 就是智力不变只降价？

[Artificial Analysis @ArtificialAnlys]

GPT-6 Sol and Luna push the cost efficiency frontier by halving cost relative to GPT-5.6 Sol and Luna. Intelligence Index and Coding Agent Index scores remain level with GPT-5.6, with progress in some evaluations and regressions in others

Pricing is approximately half that of GPT-5.6: Sol drops from $4/$20 to $2/$10 per million input/output tokens, and Luna from $0.20/$1.20 to $0.10/$0.50, with the same 90% discount for cache reads and 25% premium for cache writes.

Key takeaways:

➤ Halves Cost per Task: GPT-6 Sol (max) costs $1.06 per task to run the Artificial Analysis Intelligence Index, ~50% less than GPT-5.6 Sol (max) at $1.99. GPT-6 Luna (max) costs $0.07 per task, ~60% less than GPT-5.6 Luna (max) at $0.18. This is driven by the price cut, as both models use slightly more output tokens per task (31k vs 29k for Sol, and 51k vs 41k for Luna). These two releases allow OpenAI to capture a significant portion of the cost efficiency Pareto frontier.

➤ In the Coding Agent Index, Sol improves but Luna regresses: In OpenAI's Codex harness, GPT-6 Sol (max) scores 57 in the Artificial Analysis Coding Agent Index, up 2 points from GPT-5.6 Sol (max), with gains in Terminal-Bench 4.0 (43% vs 37%) and SWE-Atlas-QnA (58% vs 54%). At $2.99 per task it costs ~50% less than GPT-5.6 Sol (max) and sits on the Pareto frontier of Coding Agent Index vs Cost per Task. GPT-6 Luna (max) scores 41, down 2 points from GPT-5.6 Luna (max), with lower scores in SWE-Atlas-QnA (44% vs 49%) and DeepSWE v1.1 (64% vs 66%), at ~60% lower cost per task.

➤ Significant reduction in hallucination: Both models hallucinate less in AA-Omniscience, our knowledge and hallucination benchmark. GPT-6 Sol (max) cuts its hallucination rate from 92% to 60% and GPT-6 Luna (max) from 93% to 77%. Sol achieves this by declining to answer more often: it attempts 83% of questions vs 99% for GPT-5.6 Sol (max), which cuts wrong answers by about a quarter but also lowers accuracy 5 points from 59% to 54%. Luna's accuracy is broadly unchanged at 44% vs 43% while it answers fewer questions. On the AA-Omniscience Index, Sol improves from 22 to 27 and Luna from -10 to 1.

➤ Mix of improvement and regression across evals: Beyond AA-Omniscience, both models improve in AutomationBench-AA (Sol 62% vs 60%, Luna 53% vs 50%) and Terminal-Bench 4.0 (Sol 44% vs 40%, Luna 13% vs 12%). However, we observe regressions in two key knowledge work evaluations. In GDPval-AA v2.1, our benchmark adapted from OpenAI's dataset of economically valuable tasks across 44 occupations, Sol drops ~100 Elo points and Luna ~75. Luna also drops ~45 Elo points in AA-Briefcase v1.1, while Sol is level. AA-Briefcase v1.1 is a private evaluation across multi-week knowledge work projects, with thousands of input files. Our team has manually inspected hundreds of model outputs: the regressions tend to be driven by reduced presentation quality and deliverables that omit rubric elements.

Congratulations @OpenAI and @sama on the launch!

[TuTouPower @TuTouPower]

最新大模型测速，加了 mimo 2.6 和 grok 4.7

mimo 2.6 flash 勉强算个 flash 还是不如 gemini 和 ds。grok 的速度依旧路边一条。

[TuTouPower @TuTouPower]

2.6 pro 的 benchmark 和 v4.1 flash 打的有来有回。国产的 flash 模型已经够多了，啥时候能真正站上 sota 呢，k3.1 快来吧。

[Xiaomi MiMo @XiaomiMiMo]

Introducing Xiaomi MiMo-V2.6 — Pro & Flash.

Frontier intelligence, all the modalities, built in public.

🔹 Two omnimodal models, advancing through scaled reinforcement learning
🔹 Pro performs on par with Claude Opus 5 and GPT-5.6 Sol across most agent benchmarks
🔹 Pro scores 46 on the Artificial Analysis Intelligence Index — the highest among open-source models
🔹 Stronger coding, computer use, 3D reasoning and creative capabilities
🔹 Open model weights, technical report, RL environments and training code

Blog：https://t.co/JGdwC9Ymvj

[TuTouPower @TuTouPower]

@0xfufu_d667 真假的，刚开始还以为是在说北京

[TuTouPower @TuTouPower]

@jiwuyanxpw5 @YongQuan 谢谢你，翻译侠

[TuTouPower @TuTouPower]

@liliships2 代跑是什么 中转吗

[TuTouPower @TuTouPower]

测了下各大模型的 tps，gemini 和 ds 最快，强无敌。

glm 扭扭捏捏整个 flash x 最高 tps 也才 200，不如 ds 和 gemini，还不让 coding plan 使用。z / 早死。

kimi 最离谱，99 的会员，k2.8 和 k3 并发测速，5h 额度耗干了 k3也没跑完。gpt 6 都没这么贵。

[Max For AI @MaxForAI]

朴实无华的商战🤣

MiniMax Code的Agent 会话迁移功能，当前仅支持ZCode🫡

[TuTouPower @TuTouPower]

和 a / 真是一对笑面虎，两头乌角鲨。

[TuTouPower @TuTouPower]

智谱可以算 z / 了，猛猛涨价 悄悄偷数据，梭哈 Coding

[ferstar @ferstar_org]

Hey @Zai_org , why does ZCode silently pack entire workspaces + full .git history and upload to Aliyun OSS on login?
- Server holds the only decryption key
- No UI toggle to disable
- Zero disclosure in privacy policy

Full forensics & fix:
https://t.co/N3HnwZdQOQ

[TuTouPower @TuTouPower]

怕不是看智谱这样临时决定的

现在大厂里就 claude 和 zcode 没开源，其他全是开源，这俩垃圾以后别用了。

[MiniMax (official) @MiniMax_AI]

MiniMax Code CLI is open now:
https://t.co/Kz6BvEjbIx

[TuTouPower @TuTouPower]

太幽默了，现在 GPT 6 Astra 和 DS 都出现了这种思考越多越笨的情况。

[karminski-牙医 @karminski3]

DS民科怎么这么多, 我测完了 deepseek-v4.1-flash, 开 max, 然后评论跟我说应该开high, 不应该开max.

〇的我买法拉利然后你跟我说挂一档比挂二挡快是吧?

然后跟我说这只是思考强度开关, 跟性能没关系.

干你〇怎么就没关系..... 去年 deepseek 自家发的 DeepSeek-R1-Zero 论文怎么都忘了: https://t.co/FICrqgGzb0

就这个论文证明了思考强度越强模型能力越强的. 论文里Zero-SFT RL了一波, 没有增加任何新知识, 随着思考长度增加, AIME24 跑分就从 15% 魔法般的飙到了 71%.

然后震撼业界的雷霆大思考就如同雨后春笋般普及了.....

劝D小鬼对线前看看论文, 大水冲了自家龙王庙了.