Sources來源
Where a number came from.一個數字從哪來。
This edition was closed on 1 October 2026. Index scores in the model table were copied only from Artificial Analysis articles read for this desk. News headlines that sit behind a paywall are linked as a news search, not as a fabricated URL. If a figure is not on this page, do not treat it as checked.這一版在 2026 年 10 月 1 日收束。模型表的指數只從本台讀過的 Artificial Analysis 文章抄來。付費牆後面的新聞標題連到新聞搜尋,不是編造的網址。數字若不在這一頁,就不要當成已核對。
Scores and prices分數與價格
01- Claude Opus 5.5 — index 58 at max, $4 / $20 per million, cache reads $0.20, 1M context. 22 September 2026.max 指數 58,每百萬 $4 / $20,快取讀取 $0.20,1M 上下文。2026 年 9 月 22 日。
- Claude Sonnet 5.5 — index 56 at max, two points behind Opus, $2 / $10 sticker, high output volume. 28 September 2026.max 指數 56,落後 Opus 兩分,$2 / $10 牌價,輸出量大。2026 年 9 月 28 日。
- Gemini 4 Argon — index 53 at high, tied with GPT-6 Astra (max, 53), one point ahead of GPT-6.1 Sol (max, 52). Discounted $2 / $10 on a $4 / $20 list. Limited rollout. Gemini 3.8 Flash (high) is stated as 12 points lower, which is the 41 in the table. 30 September 2026.high 指數 53,與 GPT-6 Astra(max,53)持平,領先 GPT-6.1 Sol(max,52)一分。$4 / $20 牌價上的折扣價 $2 / $10。開放有限。Gemini 3.8 Flash(high)被寫成低 12 分,也就是表上的 41。2026 年 9 月 30 日。
- GPT-6.1 Sol — near-Astra index at under a quarter of Astra's cost per task. 29 September 2026.指數接近 Astra,單次任務成本不到 Astra 的四分之一。2026 年 9 月 29 日。
- AA-AgentPerf-Local — laptop and workstation agent speed, 29 September 2026. Active parameter count dominated. RTX 5090 led when the model fit in 32 GB.筆電與工作站的智能體速度,2026 年 9 月 29 日。活躍參數量主導。模型放進 32 GB 時,RTX 5090 領先。
- Artificial Analysis — the live board. Use it the day you buy.即時榜。你買的那一天用它。
- OpenAI pricing and與 Anthropic pricing — confirm stickers. Luna's $0.10 / $0.50 is September reporting, not a fresh scrape of the price page.確認牌價。Luna 的 $0.10 / $0.50 是九月報導,不是價格頁的新抓取。
Grok 4.7, Muse Spark, Qwen3.8 Max, Kimi K3, GLM-5.3, DeepSeek V4.1 Flash, MiMo, and Mistral Medium 3.5 are in the table because they are the models a buyer will be offered. Their index cells are blank on purpose. Secondary leaderboards disagreed with each other in the same week, so those ranks were not copied.Grok 4.7、Muse Spark、Qwen3.8 Max、Kimi K3、GLM-5.3、DeepSeek V4.1 Flash、MiMo 與 Mistral Medium 3.5 在表上,是因為買方會被推這些模型。它們的指數格故意空白。同一週的次級排行榜彼此不合,所以那些名次沒有抄。
Labs實驗室
02- OpenAI — GPT-6 family, Codex, Dots, GPT Image.GPT-6 家族、Codex、Dots、GPT Image。
- Anthropic — Claude 5 family, Claude Code.Claude 5 家族、Claude Code。
- Google DeepMind — Gemini, Veo, Nano Banana.Gemini、Veo、Nano Banana。
- xAI — Grok, Grok Imagine.Grok、Grok Imagine。
- Meta — Muse and the Llama open-weight line. They are different products.Muse 與 Llama 開放權重線。它們是不同產品。
- Alibaba Qwen, DeepSeek, Kimi, Z.ai — Chinese frontier and open weights.中國前沿與開放權重。
- Mistral — European open weights.歐洲開放權重。
- Black Forest Labs — FLUX.
Agents智能體
03- NousResearch/hermes-agent
- openclaw/openclaw
- Aider and與 Cline
- Architecture comparisons read while writing the agents page included OpenClaw's docs, Hermes' docs, and neutral writeups from the last week of September 2026. Characterizations on this site follow the public design each project claims: a learning loop versus a gateway. Confirm details in the repo before you install. Both projects' security advisories are on their GitHub security tabs.寫智能體頁時讀過的架構比較,包括 OpenClaw 的文件、Hermes 的文件,以及 2026 年 9 月最後一週的中立文章。本站的描述跟隨各專案公開主張的設計:學習迴圈對上閘道。安裝前到儲存庫確認細節。兩個專案的安全通報都在它們的 GitHub 安全分頁。
Local deployment本機部署
04- Ollama, LM Studio, llama.cpp, MLX.
- vLLM, SGLang, Open WebUI.
- ComfyUI for local image and video graphs.用於本機圖像與影片圖。
- Fit estimates in the advisor are Q4-class weight sizes plus a small margin, rounded to families. They are planning numbers, not a benchmark certificate. Context length eats the margin.顧問裡的容量估計是 Q4 等級的權重大小加上一點餘裕,再四捨五入到家族。它們是規劃數字,不是基準證書。上下文長度會吃掉餘裕。
News and policy, last week of September 2026新聞與政策,2026 年 9 月最後一週
05- FTC inquiry into frontier labs — Reuters, AP, New York Times, 30 September.FTC 對前沿實驗室的調查——路透、美聯社、紐約時報,9 月 30 日。
- White House meeting and voluntary frontier accord — 29 September reporting.白宮會議與自願前沿協議——9 月 29 日報導。
- California worker-protection signatures — 30 September.加州勞工保護簽署——9 月 30 日。
- Anthropic IPO prospectus coverage — Reuters and the Los Angeles Times.Anthropic 上市說明書報導——路透與洛杉磯時報。
- Bank of England comments on AI-linked debt — Financial Times, 30 September.英格蘭銀行對 AI 相關債務的評論——金融時報,9 月 30 日。
- NIST post-quantum cryptography — ML-KEM, ML-DSA, SLH-DSA. Primary source for the quantum section.ML-KEM、ML-DSA、SLH-DSA。量子一節的主要來源。
The live headline strip is Google News RSS, fetched by news.php on Hostinger and cached for about 15 minutes. If that request fails, the ten-story briefing remains. That briefing is data/news.json.即時標題條是 Google 新聞 RSS,由 Hostinger 上的 news.php 抓取,快取大約 15 分鐘。若該請求失敗,十則簡報仍在。簡報是 data/news.json。
How to refresh this desk怎麼更新這張參考台
06- Change a model in
data/models.jsonor a story indata/news.json.改data/models.json裡的模型,或data/news.json裡的故事。 - Put a new
aaIndexonly after you have opened the Artificial Analysis article and read the effort level.只有在你打開 Artificial Analysis 文章並讀過努力等級之後,才填新的aaIndex。 - Commit and push with GitHub Desktop. Hostinger deploys the branch you connected. Steps are in the README.用 GitHub Desktop 提交並推送。Hostinger 會部署你接上的分支。步驟在 README。