English translation is being prepared.
Matched fieldsBody
←Search results
English text is being prepared.
Meta releases Muse Wonder and Muse Wonder Full, surpassing Claude on long-horizon coding tests…
Meta Superintelligence Labs launched its new-generation flagship models Muse Wonder and Muse Wonder Full on October 13. The Full version scored 56 on the Artificial Analysis Composite Intelligence Index, 2 points behind Claude Opus 5.5's 58. On the DeepSWE 1.1 long-horizon software engineering test, Wonder Full scored 79.0, surpassing Opus 5.5's 74.2 and GPT-6 Astra's 74.1, but it still trails Opus on Terminal-Bench 4.0 and ProgramBench. Artificial Analysis estimates its average task cost at $4.52, up 183% from the previous generation Muse Spark 1.3; a developer comparison on the same prompt shows its 3D game generation is noticeably worse than Opus, and Meta's model card has acknowledged 3D tasks as a known limitation.
Meta Superintelligence Labs launched its new-generation flagship models Muse Wonder and Muse Wonder Full on October 13. The Full version scored 56 on the Artificial Analysis Composite Intelligence Index, 2 points behind Claude Opus 5.5's 58. On the DeepSWE 1.1 long-horizon software engineering test, Wonder Full scored 79.0, surpassing Opus 5.5's 74.2 and GPT-6 Astra's 74.1, but it still trails Opus on Terminal-Bench 4.0 and ProgramBench. Artificial Analysis estimates its average task cost at $4.
Matched fieldsBody
Meta releases the Muse Wonder series of models, focused on personal agent memory and long-horizon tasks
Meta released two models, Muse Wonder and Muse Wonder Full, internally codenamed Watermelon, targeting personal and professional agent tasks, with a focus on improving cross-conversation memory and long-duration task execution. Artificial Analysis evaluations show that Wonder Full placed in the top three on six of ten tests, and at the highest reasoning intensity, Muse Code paired with Wonder Full matches Claude Code paired with Sonnet 5.5 in coding performance. API pricing is $4 per million input tokens and $20 per million output tokens; Wonder Full is available only to Muse Code subscribers, and Meta AI head Alexandr Wang acknowledged that its coding capability still lags behind Opus 5.5.
Meta released two models, Muse Wonder and Muse Wonder Full, internally codenamed Watermelon, targeting personal and professional agent tasks, with a focus on improving cross-conversation memory and long-duration task execution. Artificial Analysis evaluations show that Wonder Full placed in the top three on six of ten tests, and at the highest reasoning intensity, Muse Code paired with Wonder Full matches Claude Code paired with Sonnet 5.5 in coding performance. API pricing is $4 per million inp
Matched fieldsBody
Ai2 Releases MolmoMotion, a Language-Guided 3D Motion Prediction Model
Ai2 releases MolmoMotion: by inputting video frames, 3D points on objects, and action descriptions, it can predict the trajectories of these points in 3D space over the next few seconds, outperforming existing prediction methods. It also open-sources the MolmoMotion-1M dataset (derived from 1.16M videos) and the PointMotionBench benchmark (2.7K video clips), and releases the model weights. In simulation, a control policy based on MolmoMotion achieved a 76.3% success rate on pick-and-place tasks, higher than the 56.0% achieved by one based on Molmo 2.
Ai2 releases MolmoMotion: by inputting video frames, 3D points on objects, and action descriptions, it can predict the trajectories of these points in 3D space over the next few seconds, outperforming existing prediction methods. It also open-sources the MolmoMotion-1M dataset (derived from 1.16M videos) and the PointMotionBench benchmark (2.7K video clips), and releases the model weights. In simulation, a control policy based on MolmoMotion achieved a 76.3% success rate on pick-and-place tasks,
Matched fieldsBody
Anthropic releases Claude Sonnet 5.5, 30% faster than Sonnet 5 with unchanged per-token pricing
Anthropic has released Claude Sonnet 5.5, the second model in the Claude 5.5 family. It is smarter and more efficient than Sonnet 5 and 30% faster, with unchanged per-token pricing; because it needs fewer tokens to complete the same work, costs for most tasks drop by up to 30%. The official recommendation is to use Sonnet 5.5 for well-scoped everyday tasks such as bug fixing and rapid feature iteration, while complex long-running work should still use Opus 5.5; Haiku 5.5 will join the family in the coming weeks. The model ID is claude-sonnet-5-5, and it is already available on platforms including the Claude API, Amazon Bedrock, Google Cloud, and Microsoft Foundry. Starting with v2.1.284, the sonnet alias in Claude Code points to this model, which runs at medium effort by default and natively supports a 1M context window.
Anthropic has released Claude Sonnet 5.5, the second model in the Claude 5.5 family. It is smarter and more efficient than Sonnet 5 and 30% faster, with unchanged per-token pricing; because it needs fewer tokens to complete the same work, costs for most tasks drop by up to 30%. The official recommendation is to use Sonnet 5.5 for well-scoped everyday tasks such as bug fixing and rapid feature iteration, while complex long-running work should still use Opus 5.5; Haiku 5.5 will join the family in
Matched fieldsBody
Fireworks Research Releases Ember-1: A Token-Saving Specialized Model Based on Kimi K3
Fireworks Research has released Ember-1, a specialized model trained on Kimi K3 that reduces token consumption by about 40% while maintaining equivalent quality. The company says that across 7 benchmarks and production traffic from two customers, K3's reasoning process can be shortened by 35–50% without any drop in accuracy; on Bedside Bench, Ember-1 sits on the Pareto frontier for cost and task completion. The model is available on Serverless as a Research Preview, offered alongside the base Kimi K3, with training support on enterprises' own data also opened up.
Fireworks Research has released Ember-1, a specialized model trained on Kimi K3 that reduces token consumption by about 40% while maintaining equivalent quality. The company says that across 7 benchmarks and production traffic from two customers, K3's reasoning process can be shortened by 35–50% without any drop in accuracy; on Bedside Bench, Ember-1 sits on the Pareto frontier for cost and task completion. The model is available on Serverless as a Research Preview, offered alongside the base Ki
Matched fieldsBody
English translation is being prepared.
Matched fieldsBody
GPT-6 Family's First Model Astra Released, Author's Hands-On Test Consumed 4 Billion Tokens
OpenAI released Astra, which the author calls the beginning of the GPT-6 family; it performs outstandingly on ARC-AGI-3 and scored highest on Zapier's AutomationBench, with pricing the same as Fable 5.1. During weekend trials, the author consumed 4 billion tokens and found its capability fluctuates noticeably—it is a showcase model rather than one that reliably does work. Astra in Codex can now continue advancing tasks without waiting for an answer, and OpenAI also claims it has achieved its automated research intern goal, with the next goal being an automated AI researcher by March 2028.
OpenAI released Astra, which the author calls the beginning of the GPT-6 family; it performs outstandingly on ARC-AGI-3 and scored highest on Zapier's AutomationBench, with pricing the same as Fable 5.1. During weekend trials, the author consumed 4 billion tokens and found its capability fluctuates noticeably—it is a showcase model rather than one that reliably does work. Astra in Codex can now continue advancing tasks without waiting for an answer, and OpenAI also claims it has achieved its aut
Matched fieldsBody
English translation is being prepared.
Matched fieldsBody
Google DeepMind releases Gemini 3.8 Flash and 3.8 Flash Cyber
Google DeepMind has released Gemini 3.8 Flash and Gemini 3.8 Flash Cyber, aimed at trusted defenders, calling them its most capable reasoning and coding models to date, with speed and cost comparable to 3.7 Flash. 3.8 Flash surpasses most larger frontier models on DeepSWE v1.1, scores 54.9% on HLE-Verified, and performs more reasoning steps and iteratively calls tools on complex tasks. 3.8 Flash Cyber reaches frontier-level performance on CyberGym, with an internal vulnerability-discovery benchmark success rate exceeding 70% across 20 programming languages, and a CWE-Bench patch pass@1 of 47.2%.
Google DeepMind has released Gemini 3.8 Flash and Gemini 3.8 Flash Cyber, aimed at trusted defenders, calling them its most capable reasoning and coding models to date, with speed and cost comparable to 3.7 Flash. 3.8 Flash surpasses most larger frontier models on DeepSWE v1.1, scores 54.9% on HLE-Verified, and performs more reasoning steps and iteratively calls tools on complex tasks. 3.8 Flash Cyber reaches frontier-level performance on CyberGym, with an internal vulnerability-discovery benchm
Matched fieldsBody
Meta's Muse Glimmer 30B Launches on Fireworks, Targeting Persistent Agents
Fireworks has launched Meta's Muse Glimmer 30B dense model under the Apache 2.0 license, supporting 128K+ token context and native image understanding, targeting persistent agents that require multi-turn tool calling. The model scores 75.5 on MCP Atlas, 74.6 on DeepSearch QA, and 51.2 on SWE-Bench Pro, all higher than the corresponding results of Gemma 4 31B and Qwen 3.6 27B; these benchmark figures are reported by Meta. Fireworks also provides recommended settings of temperature 1.0, top_p 0.95, and top_k 64, and supports adjusting reasoning intensity in the system prompt.
Fireworks has launched Meta's Muse Glimmer 30B dense model under the Apache 2.0 license, supporting 128K+ token context and native image understanding, targeting persistent agents that require multi-turn tool calling. The model scores 75.5 on MCP Atlas, 74.6 on DeepSearch QA, and 51.2 on SWE-Bench Pro, all higher than the corresponding results of Gemma 4 31B and Qwen 3.6 27B; these benchmark figures are reported by Meta. Fireworks also provides recommended settings of temperature 1.0, top p 0.95
Matched fieldsBody
English translation is being prepared.
Matched fieldsBody
Ethan Mollick Tries GPT-5.5 Early: Capability Gains Are Still Accelerating
Ethan Mollick was granted early access to try GPT-5.5 and believes its capabilities have improved markedly; GPT-5.5 Pro completed the same 3D port town simulation task in 20 minutes, compared with 33 minutes for the previous generation, GPT-5.4 Pro. Using Codex together with GPT-5.5 to process a decade's backlog of his own crowdfunding data, he generated a paper including a genuine literature review and statistical methods with just four prompts, and he considers its quality equivalent to a second-year PhD project. He also noted that OpenAI's new image model (sometimes called GPT-imagegen-2) can render high-quality text and images, but that AI still has rough edges in long-form fiction writing, such as florid prose and homogeneous dialogue.
Ethan Mollick was granted early access to try GPT-5.5 and believes its capabilities have improved markedly; GPT-5.5 Pro completed the same 3D port town simulation task in 20 minutes, compared with 33 minutes for the previous generation, GPT-5.4 Pro. Using Codex together with GPT-5.5 to process a decade's backlog of his own crowdfunding data, he generated a paper including a genuine literature review and statistical methods with just four prompts, and he considers its quality equivalent to a seco
Matched fieldsBody
English translation is being prepared.
Matched fieldsBody
English translation is being prepared.
Matched fieldsBody