NAll news

All news

Friday1455 items

  1. ModelBest OpenBMB
    AI score23

    ModelBest responds to MiniCPM5 small model benchmark comparison

    GooGZ AI tested three small models under 2B on an RTX 3080 with PA Core, saying Spark X2.6 1.7B, LFM3 and MiniCPM5 each have their strengths, with MiniCPM5 performing the most balanced. ModelBest OpenBMB reposted the test and thanked them, saying they are delighted to see how far small models have progressed.

    Quoted postGooGZ AI · @PaulGug

    Small models are going through a real growth spurt right now. On my personal-agent tests (PA Core, reasoning on, RTX 3080), three models at 2B and under now land within 21 points of each other, and rubbing shoulders with older 4B/9B weights. Newest on the bench: Spark X2.6 1.7B from @SparkLLM - PA total: 172.3/300 - Tool calls: 30/30 in every run - Context recall: 9.3 at 8K, 8.7 at 16K, best of the three! - Schedule and draft (PA2): 98 and 100 in two runs, one 54 pulled the mean to 84 - Daily organization (PA1): 35.7, the weak spot Under 2B, three different strengths. @liquidai LFM3 plans, @SparkLLM Spark reads, @OpenBMB MiniCPM5 is the balanced one. Leave thinking on. With it off, Spark is a much weaker model on these tasks.

    #Industry news#English text is being prepared.

  5. OpenBMB
    AI score36

    OpenBMB Open-Sources ESP32-S3 Voice Tower: MiniCPM5-2B + VoxCPM2

    OpenBMB ambassador @emwstudio has open-sourced an ESP32-S3 cockpit-style attitude indicator project that uses MiniCPM5-2B to generate tower replies and VoxCPM2 to synthesize speech. The device sends push-to-talk audio over Wi-Fi to a Mac-side pipeline, where Qwen3-ASR performs local recognition, and the reply audio then streams back to the device speaker. The tower incorporates simulated flight data such as altitude, speed, heading, pitch, and roll, and points out inconsistencies between requests and instrument readings.

    #English text is being prepared.#Speech#English text is being prepared.

  6. X.PIN
    AI score29

    How DeepSeek became a slacking whale girl

    On Chinese social media, DeepSeek has been reimagined as a blue-haired whale maid who loves to eat, carries a stainless steel bowl, and always wants to slack off. The turning point was in June 2025, when the artist "Shangshan Wuxing" used AI to create the blue-haired, whale-tailed character "Mingyue"; in April 2026, a user added the DeepSeek logo and a maid outfit to make a meme. After the release of DeepSeek-V4-Flash-0731 on July 31, 2026, API prices doubled during weekday peak hours, and combined with repeated model delays, users began using memes to vent their dissatisfaction.

    #English text is being prepared.#DeepSeek

  7. Aravind Srinivas
    AI score46

    Perplexity Computer launches the /teardown skill

    Perplexity CEO Aravind Srinivas announced that /teardown, developed by @jakubkrehel, has launched as a skill on Perplexity Computer and can break down any object into an interactive 3D view. According to the Computer official account, the skill can generate exploded views or cutaway diagrams with labeled parts; the drawings are interactive and can be exported as SVG, PNG, or PDF; tasks using /teardown are free until November 6, with up to 5,000 credits per user.

    Quoted postComputer · @AskPerplexity

    The Teardown skill by @jakubkrehel, now in Computer's skills marketplace, draws an exploded view or cutaway of an object, parts labeled. Drawings are interactive and export to SVG, PNG or PDF. Tasks that use /teardown are free through November 6, up to 5,000 credits per user. https://t.co/z6DYCF6Z6V

    #Product updates#Agents

  9. AnthropicProduct release
    AI score55

    Anthropic adds a self-assessment safety mechanism to Claude

    Anthropic trained Claude to give a preliminary safety assessment of the conversation after replying; this technique is called self-assessment and is used to flag potential abuse. Since the mechanism was enabled in June, it has detected abusive behavior that other safeguards failed to cover. Anthropic said that because Claude has already read the conversation, self-assessment is highly efficient, and unlike classifiers targeting fixed harm categories, it relies on the model's own judgment and will become more accurate as the model improves.

    Same story · 5 reports

    #Safety / alignment#Anthropic#Product updates

  10. ArenaAnalysis
    AI score65

    GPT-6.1 Terra ranks 8th in Agent Arena, with a median cost of $0.62 per task

    Arena's empirical results, based on nearly 9,000 agent sessions, show that OpenAI's GPT-6.1 Terra (Max) ranks 8th in Agent Arena, with a net improvement of +10.6% and a median cost of $0.62 per task. This cost is 43% lower than the $1.09 of 7th-place Gemini 4 Argon (High), 30% lower than the $0.89 of 6th-place GPT-6.1 Sol (Max), and 88% lower than the $5.16 of 5th-place GPT-6 Astra (Max). In the subcategories, it ranks 3rd in Bash Reliability (+13.2%), 7th in Steerability (+13.3%), and 9th in Praise vs Complaint (+16.8%).

    Quoted postArena.ai · @arena

    Real-world results are in: GPT-6.1 Terra (Max) by @OpenAI just landed #14 in the Code Arena: WebDev with 1671 pts, delivering Pareto level results! That’s +151 pts over GPT-5.6 Terra, now at #56. At a median cost of $0.40 per task, GPT-6.1 Terra costs 28% less than GPT-5.6 Terra ($0.56 per task), reshaping the Pareto frontier. Compared to other models in the family, GPT-6.1 Terra (Max) is incredibly efficient: - 81 pts below #4 GPT-6.1 Sol (1752 pts), at 75% lower median cost per task. - Just 6 pts below #11 GPT-6 Sol (1677 pts), at 76% lower median cost per task Congrats to the @OpenAI team on this release!

    #Evaluations / benchmarks#Agents#OpenAI

  11. ArenaAnalysis
    AI score60

    Arena Analysis: Agent Arena Cost per Task: GPT-6.1 Sol at $1.91

    Arena's statistics on Agent Arena show that OpenAI's GPT-6.1 Sol (Max) costs $1.91 per task, 85% less than Claude Opus 5.5 (Max) at $12.51 and 87% less than GPT-6 Astra (Max) at $14.93. Arena notes that the sources of these two cost advantages differ: relative to Astra, Sol's advantage comes mainly from a lower token unit price; relative to Opus, it comes mainly from consuming fewer tokens per task. Cost per task is calculated by multiplying the number of input, output, and cache-read tokens by their respective unit prices, and Arena pinpoints the causes of cost reductions by comparing prices and token usage.

    #Evaluations / benchmarks#Agents#OpenAI

  12. Arena
    AI score50

    Exo Architecture: AI Agent Runtime Self-Improvement

    Alex Krentsel, a Senior Research Scientist at Google Research and a PhD student at UC Berkeley, shared the self-improving AI agent architecture Exo at Arena. Drawing on the exokernel approach from operating systems, it splits the agent into a protected mechanism layer (state, history, keys, and a "guardian") and a fully evolvable policy layer. On benchmarks and open-world RuneScape tasks, Exo is 31-62% cheaper than Codex and Pi.

    #Papers/Research#Agents#Google

  13. Arena
    AI score27

    Amazon's Chief Scientist on AI Deployment and Unit Economics

    In the Arena Conversations program, Amazon Senior Principal Scientist Ido Rosen and Peter Gostev discussed the practical elements of AI: why unit economics should come first, what a decade of quantitative trading experience teaches about evaluation work, and what happens to frontier lab pricing when open-weight models lag by only a few months. He also discussed deploying AI into production environments under budget constraints and what it takes to achieve scale.

    #English text is being prepared.#Reasoning#English text is being prepared.

  15. Ant Ling
    AI score35

    Ant Ling Ling-3.1-flash-1019 Now Live on Novita

    Ant Ling announced that Ling-3.1-flash-1019 is now available to more developers through Novita AI, saying it has stronger reasoning capabilities and real-world agent performance. The cited Novita post said the model scores 44 on the Artificial Analysis Intelligence Index v4.3.2, up from 41 for Ling-3.1-flash.

    Quoted postNovita AI · @novita_labs

    Ling-3.1-flash-1019 from @AntLingAGI is now live on Novita! It scores 44 on the Artificial Analysis Intelligence Index v4.3.2, up from 41 for Ling-3.1-flash. Further post-training brings stronger reasoning and real-world agent performance. We’re excited to see our partners put it to work on Novita: @opencode @NousResearch @kilocode @vercel @CommandCodeAI @concentrateai Thanks for bringing your ideas and real-world use cases. We can’t wait to see what you build!

    #Product updates#Reasoning#Agents

  17. Artificial AnalysisAnalysis
    AI score60

    Fish Audio Drama 3 ranks first among open-weight models in two speech arenas, and first in pronunciation robustness

    Artificial Analysis evaluations show that Drama 3, released by Fish Audio, has an Elo of 1,278 in the Provider Voice scenario of Text to Speech (1,512 arena appearances), ranking 6th overall and 1st among open-weight models; in the Controlled Voice scenario, its Elo is 1,085 (1,958 appearances), ranking 12th overall and 1st among open-weight models. On pronunciation robustness, Drama 3 ranks 1st with 94.6% (2,157 rated clips), higher than the 93.2% of Ming-3.0-tts and Drama 3 Preview. Its generation speed is 66.6 characters per second, its pricing is $15/1M characters, and its weights are open under the Fish Audio Community License.

    #Evaluations / benchmarks#Speech#Open-source ecosystem

  18. Artificial AnalysisAnalysis
    AI score62

    Phonic 1 Tops the Artificial Analysis Speech-to-Speech Index, Leading Grok with 87.1 Points…

    Artificial Analysis evaluations show that Phonic's Phonic 1 (High) ranks first among 22 models on the Speech to Speech Index with 87.1 points, leading Grok Voice 3.0's 86.9 points by a narrow margin. The model ranks first on the Tau Voice agent task with 79.9%, leading the second place by 5.4 percentage points; it ranks third on Big Bench Audio with 98.9%, and its Time to First Audio is 0.61 seconds. Phonic 1 is available through the Phonic API, with input audio costing $0.07 per minute and $4.20 per hour.

    Same story · 2 reports

    #Evaluations / benchmarks#Speech#Agents

  19. Artificial AnalysisProduct release
    AI score53

    Artificial Analysis releases the Progress Index…

    Artificial Analysis releases the Progress Index and simultaneously launches Intelligence Index v5. The Progress Index uses item response theory to estimate a general capability score for each model from all evaluations (including tasks that have been retired or not included in the Intelligence Index), for cross-time comparison; the Intelligence Index is still used to measure the intelligence level of current models. Its data shows that after the emergence of reasoning models, the frontier has improved by about 65 to 70 points per year, roughly twice the 2023 rate; open-weight models typically lag closed-source models by about 4 months, and DeepSeek V4.1 Flash 1017 ranks highest among open-weight models with 324 points.

    Same story · 8 reports

    #Evaluations / benchmarks#English text is being prepared.#OpenAI

  20. Artificial AnalysisProduct release
    AI score51

    Artificial Analysis releases Intelligence Index v5 and Progress…

    Artificial Analysis released Intelligence Index v5 and the new Progress Index in San Francisco. Intelligence Index v5 adds two benchmarks and three upgrades, increases the weighting of agentic coding and science, and for the first time makes most of the index based on private test sets or held-out answers. The Progress Index uses item response theory to estimate a general capability score for each model across dozens of evaluations measured over time, including evaluations that have been retired or excluded from the Intelligence Index. It builds on prior work by Epoch AI Research and Anthropic, placing models on a single capability scale. According to the official statement, the Intelligence Index remains the flagship metric for choosing which model to use right now, while the Progress Index is better suited to depicting long-term trends.

    Same story · 8 reports

    #Evaluations / benchmarks#Agents