EQX (@eqxd115) header
EQX (@eqxd115) profile picture
Joined November 2025
Photo by EQX (@eqxd115)Photo by EQX (@eqxd115)Photo by EQX (@eqxd115)Photo by EQX (@eqxd115)Photo by EQX (@eqxd115)Photo by EQX (@eqxd115)
Cuanto abran pagado Anthropic y OpenAI para dejar mal a los modelos OSS?
Announcing Artificial Analysis Intelligence Index v5, featuring 2 new benchmarks and 3 upgrades. For the first time, the majority of the Index is built on private datasets The Artificial Analysis Intelligence Index is our composite measure of model capabilities across tasks reflecting real-world usage today. v5 is our biggest upgrade since v4 launched in January. With v5, 60% of the Index is now based on evaluations with private test sets or held-out answers. Alongside Intelligence Index v5, we’re launching the Artificial Analysis Progress Index v1, which tracks model capability over time on one continuous scale. New to the Intelligence Index: ➤ FrontierCode 2.0: Cognition's newly-launched private benchmark of real-world coding tasks, graded on whether a repository maintainer would merge the agent's change ➤ Terminal-Bench-Science: 70 agentic scientific research tasks, run in a terminal environment Upgraded: ➤ HLE-Diamond (replacing Humanity's Last Exam): a refined 1,000-question version of Humanity's Last Exam, split evenly between reasoning and knowledge questions ➤ AA-Omniscience v1.1: An updated dataset with saturated and public set questions removed, new questions added, and ambiguous or erroneous questions revised ➤ CritPt v2: a private, 40-question hard subset of CritPt v2, revised by the CritPt authors and run with a code-execution tool Staying in the Index: AA-Briefcase, AutomationBench-AA, GDP.pdf and Terminal-Bench 4.0 Removed from the Index: GDPval-AA v2.1, SciCode, and AA-LCR v1.1 Key results: ➤ Claude Opus 5.5 (max) leads at 58, ahead of GPT-6 Astra (max, 57) and GPT-6.1 Sol (max, 55): Anthropic and OpenAI hold the top five models. Gemini 4 Argon (high) is next at 50 ➤ The open weights gap widens from 3 points to 12: Kimi K3.1 Pro (weights not yet released) is the top open weights model at 46. Among released weights, DeepSeek V4.1 Flash 1017 (max) and GLM-5.3 (max) lead at 32 - 26 points behind the leader and below Claude Haiku 5.5 (max, 36) ➤ Cost per task varies by more than 13x among the leading models: of the 10 highest-scoring models shown, GPT-6.1 Terra (max) is the cheapest at $0.96 per Intelligence Index task and Claude Fable 5.1 (max) the most expensive at $13.04
6
Así se ha encontrado su pequeño negocio una mujer tras los disturbios en Barcelona:
88
255
26
1,813
79,832
MiniMax M3.1 Flash: Space Performance + Bunny Speed 🚀🐰 Yes, Space Bunny - which has topped global leaderboards with 200T+ tokens - was an earlier version of M3.1 Flash. 💰Redefining the performance economics of agentic AI. M3.1 Flash achieves frontier-level intelligence across key tasks such as Cowork (Automation-Bench: 62%; ALE: 30%), Coding (DeepSWE: 74%) and Visual Reasoning (BabyVision: 91%), at 1/3 the cost of M3 and 3x the speed. API pricing (flat rate): $0.15 / $0.60 per 1M input/output tokens, with cache reads at $0.015 / 1M tokens. Model weights will arrive on Oct 23. API: platform.minimax.io M Plan: minimax.io/m-plan MiniMax Code for desktop: agent.minimax.io/download
474
679
696
7,487
1,731,994
Introducing Muse Wonder, our new family of frontier models with state of the art performance for personal and professional agentic work.
2,424
3,084
2,838
38,905
6,150,024
Replying to @PabloGrueso
Pues le voy a contestar, Pablo, ya que me da usted pie. ¿Sabe qué pasa? Que está usted reclamando un mundo que ya no existe. La Españita de Canal Plus: el chándal morado del Madrid para lavar el coche los domingos, el soniquete de Terminator y el acento inglés de Robinson antes del partido, el buen rollito socialliberal, el vecino que dudaba el día antes de las elecciones entre el PP y el PSOE y que decidía presidentes. Ese mundo ya no existe, Pablo. Llevamos décadas moviendo el marco de lo que se puede decir a base de validar cosas que eran impensables. Validaron el escrache como "jarabe democrático", ¿se acuerda? yo me acuerdo, ¿y ahora qué pasa? que se lo hacen a ellos y lo llaman fascismo. Validaron rodear el Congreso, y ahora una cacerolada delante de Ferraz la muelen a palos, ¿verdad? Validaron que un ministro le conteste "saco de mierda" a un ciudadano sin cargo público desde su cuenta de Twitter, y ahora cuando la gente les grita a ellos hijos de puta se quejan de que peligra la democracia. Mis muertos, la ventana la movieron ellos, Pablo, y ahora no la pueden cerrar. No pida usted que la cierren. Y algunos lo avisamos. Yo ya estaba diciendo esto en 2017: cuidado con el pendulazo, cuidado con el pendulazo, subnormales, que luego vendrán los lloros. La física es exacta, un péndulo llega por el otro lado justo hasta la altura desde la que lo soltaste. Todo lo que lo han empujado hacia la izquierda es lo que va a volver hacia la derecha. Y por el centro, que es donde usted quiere que se pare, no se va a parar. Tienes, Pablo, a la que señalaba víctimas de ETA antes de serlo gobernando, de facto, este país. Tienes a una cloaca contratando narcos para eliminar a jueces, Pablo. Tienes una ciudad autónoma literalmente invadida por un país vecino, y al gobierno señalando a sus habitantes, llamándolos fachas. Tienes una catástrofe natural, y al gobierno dejando que la gente se ahogue en el fango, sin enviar ayuda en tres días, para poder culpar al gobierno autonómico del desastre, y tratar de recuperar terreno electoral perdido. ¿Sigo? Los jóvenes, Pablo, los que ahora son jóvenes, no como usted ni como yo, que tenemos la misma edad, nacieron con ZP. Eran niños cuando el 15M, les cerraron el instituto con el Covid y les prohibieron bajar al parque, y en clase, llevan diez años explicándoles que la degradación de su barrio es cosmopolitismo e integración, que el coche en el que van al pueblo a ver a la abuela es un crimen climático, que su padre es sospechoso de maltrato. Llevamos diez años de un infotainment cuyo único producto es humillar a la extinta clase trabajadora, a la que llaman "cuñados". Una década diciéndole al niño en el colegio que es un potencial violador, y que tiene que deconstruir su peligrosa identidad. Esos chavales no han visto nunca el mundo que vimos usted y yo. No tienen nada que echar de menos. No puedes pedirles que respeten unas formas que a ellos nadie les ha aplicado jamás. Se llama descomplejización, lo trataban con mucho tino @aparachiqui y @JorgeSMiguel en Extremo Centro. —Que tu me llamas a mi fascista, y te gusta jugar a la guerra civil de mentirijillas, vale, nosotros también jugamos, y te sueltan "pesetas y cunetas" y "¿Que los nacionales fusilaron a tu abuelo? poco me parece".— Cito textualmente, tengo el libro subrayado por delante. De lo que se trata, Pablo, es de que no le puedes pedir moderación ni mesura a muchacho de 20 años, porque moderado es el que tiene algo que perder, usted y yo lo somos porque tenemos nómina, casa, hipoteca y niños en el cole. Un chaval al que le han dicho a al cara que no va a poder comprarse una casa, que esta viendo al gobierno y sus voceros decirle que construir para él una VPO es de fachas, que paga cientos de euros por un cuartucho, que su barrio es un laboratorio de integración y que si se queja es por fascista, a ese, ¿le quitamos también el derecho a enfadarse? Nosotros somos la generación que vio pasar todo esto y se quedó mirando La Sexta y leyendo El País: la que se calló lo del jarabe democrático por no parecer facha, la que miró para otro lado cuando le tiraban pis a Arrimadas en el Orgullo, cuando escracheaban a Villacís embarazada, o la que miró para otro lado cuando reconocieron a Cristina Cifuentes haciendo la compra, la rodearon, la escupieron y la insultaron delante de su hijo, que veía la escena llorando. Lo que hay hoy lo hemos hecho nosotros, con nuestras miserias y nuestros miedos, y esos chavales que no tienen mesura son la generación que va a sacar este país del fango, en Paiporta y en todas partes, y lo va a hacer con una ventana que no abrieron ellos pero que van a usar. Así que cuantas menos lecciones le demos usted y yo, Pablo, mejor.
648
6,351
803
15,560
805,205
Introducing Gemini 4 Argon – our new frontier model. It’s built for complex workflows across coding, enterprise knowledge work, and cybersecurity defense – rolling out today to a set of trusted testers through our Fairwind Program.
1,986
4,631
4,527
42,816
9,030,492
EQX retweeted
We analyzed how @claudeai’s Opus 5.5 writes compared with Opus 5 across high-reasoning Text Arena outputs. 10 of 12 writing measures moved in a better direction. Opus 5.5 should be easier to read: - Long content words fall from 41.7% to 38.6%, the lowest share of any Claude model we analyzed. - Sentences are 17% shorter on average, dropping from 12.14 to 10.03 words. The tradeoff is length. Answers get 6% wordier, rising from 453 to 481 words on average, making Opus 5.5 give the longest answers across the Opus family. It also sounds less recognizably AI on two familiar tells: - 95% fewer em dashes - 73% fewer semicolons But a new giveaway may be emerging. Hedges and caveats such as “perhaps” and “arguably” rise 97%, from 0.39 to 0.77 per 1,000 words, the highest rate of any Claude model we analyzed. What do you think: does Opus 5.5 read more naturally?
97
122
78
2,103
1,508,098
AI is getting cheaper more quickly than any other transformative tech in history. At a given level of performance, cost has fallen ~47%/quarter since 2023. That’s 4× faster than DNA sequencing, 6× faster than compute, 18× faster than lithium batteries, and (up to 1973) 54× faster than electricity.
209
1,390
337
8,835
1,981,738
Opus 5.5 performs at the level of Fable 5.1. It's ~30% faster and ~40% cheaper than Opus 5 per task. In Claude Code: - 5-hour session limits increase 20% today - Opus 5.5 is priced lower, so it goes 25% further within limits - Pro, Max, and Team users get a reset to use anytime
Introducing Claude Opus 5.5, the first model in our new Claude 5.5 family. It performs at the level of Claude Fable 5.1 for most tasks, and costs 40% less to run than Opus 5.
465
1,204
553
16,248
1,555,584
Existiendo pistolas, es descorazonador ver a una persona apuñalando a otra y a los transeúntes intentando frenarla con carteles, palos y un rodillo de amasar. Prohibamos también los cuchillos, los tenedores y los bates, a ver si así "los malos" se pasan a las guerras de almohadas
Sonntagvormittag ist es in Kirberg in Westhessen zu einem versuchten Tötungsdelikt gekommen. Ein 50-Jähriger mit deutschem Pass, geboren in Kasachstan, betritt eine Bäckerei und greift dort seine getrennt lebende Ehefrau mit einem Küchenmesser an. Vor der Bäckerei greift er erneut an und sticht auf sie ein. Sie erleidet schwere Stich- und Schnittverletzungen. Umstehende schlagen mit Nudelholz, Schild und Schaufel auf ihn ein und drängen ihn von ihr weg. nius.de/nachrichten/50-jaehr…
1
1
26
1,264
Grok 4.7 is here. It's a notable improvement over Grok 4.6 at the same price and speed.
1,607
3,130
1,938
28,889
22,017,803
Menos mal que tenemos a las empresas chinas publicando sus modelos OSS. Si fuese por Anthropic y OpenAI la IA seria algo exclusivo para ellos, el gobierno y cualquiera que pague 1000$ por millon de tokens.
We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so. Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training. You can read the full post here: darioamodei.com/post/we-must…
230
Muse Spark 1.3 is rolling out today with frontier performance almost too cheap to meter. This is the biggest jump we've made so far on coding and agentic work. Try it in Muse Code and our API. Next up 🍉 and Muse Spark open weights releases coming soon.
1,445
1,601
1,228
19,877
3,805,050
79
1,444
61
26,473
572,925
GLM-5.3-Flash can now be run locally! ✨ Run 3-bit on 128GB RAM via Unsloth GGUF. GLM-5.3-Flash (ox-alpha) rivals Claude Opus 4.8 on DeepSWE, coding & agentic benchmarks. Guide: unsloth.ai/docs/models/glm-5… GGUF: huggingface.co/unsloth/GLM-5…
Introducing GLM-5.3-Flash - Leading capabilities at a highly competitive price - Natively multimodal with a 1M-token context window - A 320B-A18B model released under the MIT License - Previously previewed as Ox Alpha, running entirely on Chinese AI chips Blog: z.ai/blog/glm-5.3-flash Available now across all official platforms: Weights: huggingface.co/zai-org/GLM-5… API: docs.z.ai/guides/llm/glm-5.3… Coding Plan: z.ai/subscribe ZCode: zcode.z.ai/en Chat: chat.z.ai AutoClaw: autoclaw.z.ai
102
198
45
1,668
307,173
GLM-5.3-Flash scores 57 on the Artificial Analysis Intelligence Index. At $0.09 Cost per Task, it sits comfortably on the Intelligence vs. Cost per Task Pareto frontier @Zai_org has released GLM-5.3-Flash, a smaller and cheaper sibling to GLM-5.3 at 320B total parameters and just 18B active parameters. GLM-5.3-Flash supports low/high/max reasoning efforts, and scores 57 evaluation on the Artificial Analysis Intelligence Index with max reasoning effort. This places the model only 3 points behind GLM-5.3 at 60 and in line with GPT-5.6 Terra and Muse Spark 1.2. On Z AI's first-party API, GLM-5.3-Flash is priced at $0.15 / 1M input tokens and $0.50 / 1M output tokens, just over 10% of the price of GLM-5.3. Cached input tokens are priced at $0.026 / 1M tokens, an 80% discount. Its Cost per Task on the Intelligence Index is $0.09, compared to $0.68 for GLM-5.3 (max), and it sits on the Pareto frontier for Intelligence vs. Cost per Task. Key results: ➤ GLM-5.3-Flash is 3 points behind GLM-5.3 (max) on the Artificial Analysis Intelligence Index, at ~7.5x lower Cost per Task. At $0.09 per Intelligence Index task against $0.68 for GLM-5.3, it sits on the Pareto frontier for Intelligence vs. Cost per Task. It ties GPT-5.6 Terra ($0.51) and Muse Spark 1.2 ($0.40) at 57 while costing ~5.7x and ~4.4x less per task. ➤ GLM-5.3-Flash is less token efficient, but its low per-token pricing means this does not translate into a high Cost per Task. The model used 149M output tokens to run the Intelligence Index, ~11% fewer than GLM-5.3 at 168M, but more than Kimi K3 (133M) and Qwen3.8 2.4T A95B (136M) which score the same on the Intelligence Index. Reasoning tokens account for 134M of the 149M total (~90%). ➤ GLM-5.3-Flash matches GLM-5.3 on real-world agentic work on GDPval-AA v2. With an Elo of 1770, the model is tied within the margin of error for GLM-5.3 and Grok 4.6. This places it behind only Claude Opus 5 (xhigh and max). On Terminal-Bench v2.1 it also matches GLM-5.3 (84.3% vs 83.9%), and on τ³-Banking it trails by 3.1 p.p. at 47.2%. ➤ GLM-5.3-Flash demonstrates good real-world knowledge and hallucination rate, scoring +7 on AA-Omniscience. Its AA-Omniscience Accuracy is 28%, 6 p.p. below GLM-5.3 (max) at 34% and well below GPT-5.6 Terra at 47%. However, with a Hallucination Rate of 28%, it is an improvement over GLM-5.3 at 30%. In real-world knowledge, GLM-5.3-Flash knows less than the bigger models and frontier proprietary models in its Intelligence Index tier with an accuracy of 28%. Additional model details: ➤ Pricing: On Z AI's first-party API, $0.15 / 1M input tokens and $0.50 / 1M output tokens . Cached input tokens are priced at $0.03/ 1M tokens, an 80% discount. ➤ Accessibility: Accessible through Z AI's first-party API at launch. ➤ Size: 320B total parameters with 18B active parameters ➤ License: MIT ➤ Context Window: 400k
61
169
56
1,945
167,712
EQX retweeted
Introducing GLM-5.3-Flash - Leading capabilities at a highly competitive price - Natively multimodal with a 1M-token context window - A 320B-A18B model released under the MIT License - Previously previewed as Ox Alpha, running entirely on Chinese AI chips Blog: z.ai/blog/glm-5.3-flash Available now across all official platforms: Weights: huggingface.co/zai-org/GLM-5… API: docs.z.ai/guides/llm/glm-5.3… Coding Plan: z.ai/subscribe ZCode: zcode.z.ai/en Chat: chat.z.ai AutoClaw: autoclaw.z.ai
999
2,604
2,012
23,789
6,979,321
EQX retweeted
Aloha! 🌺Introducing Ornith-1.5, a family of open-source LLMs spanning 9B Dense, 35B MoE, and 397B MoE, trained with self-improving strategies. It achieves state-of-the-art performance among open-source models of comparable size and delivers performance comparable to Claude Opus 4.8 across reasoning, agentic, and coding tasks: ✅Terminal-Bench 2.1 (86.1) ✅SWE-Bench (86 on verified, 65.1 on pro, 79.6 on Multilingual) ✅DeepSWE (56) ✅HLE (44.6) ✅ClawEval (81.4) ✅Tool Decathlon (71.2) Ornith-1.5 takes a major step toward training foundation models through end-to-end self-improvement, extending the self-scaffolding strategies introduced in Ornith-1.0 into a more complete self-improvement loop: the model proposes new tasks, generates task-specific scaffolds, and produces solution rollouts for reinforcement learning, continuously creating new learning experiences from which it can improve. All models, along with their quantized versions (FP8, GGUF, MLX, and NVFP4), have been released under the MIT License, enabling unrestricted commercial and research use. 📘Tech Blog: ornith.ai/ornith_1_5.html 🤗Huggingface: huggingface.co/collections/o…
411
934
679
6,801
4,564,503
We promised open weights for Qwen3.8. Now, time to meet them! 🎉 ⚡ Qwen3.8-27B: - A native multimodal dense model. With just 27B parameters, it outperforms Qwen3.7-Plus overall and shines in real-world coding & office workflows. - 262K native context, easily extendable to 1M tokens via YaRN. - Built for builders. Highly efficient, high-quality, and licensed under Apache 2.0. 🚀 The open weights for Qwen3.8-2.4T-A95B (Max-level) have also been released recently. Whether you're shipping lightweight applications with Qwen3.8-27B locally or building agents with Qwen3.8-2.4T-A95B, they're yours now! Download, deploy, and build something we haven't imagined yet. 👀👇 - Hugging Face: huggingface.co/collections/Q… - ModelScope: modelscope.cn/collections/Qw…
849
2,112
976
15,911
6,045,939
Today we're also opening the weights for Muse Glimmer, a great 30B parameter dense model that can run locally. Soon we'll also release the weights for Muse Spark 1.2, our latest foundation model. Meta is a strong supporter of open source and I'm proud of these releases. Congrats to @alexandr_wang and the MSL team for all your great work on these models.
1,852
2,518
1,247
30,300
5,292,754