Google has released the large language model Gemini 4 Carbon and the world model Genie 4; the former surpasses its own Gemini 4 Argon and OpenAI's GPT-6 Astra on comprehensive capability benchmarks. Gemini 4 Carbon focuses on vulnerability detection and patching and is currently available only to select cybersecurity experts; Genie 4 adds audio capabilities, allowing developers to add background sound and dialogue to AI worlds, and a waitlist is now open.
Mistral has released the weights of Mistral Large 4 on Hugging Face. The model, nicknamed Le Chonk, has approximately 1 trillion parameters with 52 billion activated per step, and supports text and image inputs. On Terminal-Bench 4.0, its score rose from 28% in the preview version to 41%, slightly above GLM-5.3's 40%, but on the new version of the Artificial Analysis Intelligence Index it scored only 26 points, below the 32 points of DeepSeek V4.1 Flash and GLM-5.3. API pricing is $0.70 per million input tokens and $2.10 per million output tokens, under a modified MIT license; using it as a commercial cloud service requires separate authorization.
Mistral has released the weights of Mistral Large 4 on Hugging Face. The model, nicknamed Le Chonk, has approximately 1 trillion parameters with 52 billion activated per step, and supports text and image inputs. On Terminal-Bench 4.0, its score rose from 28% in the preview version to 41%, slightly above GLM-5.3's 40%, but on the new version of the Artificial Analysis Intelligence Index it scored only 26 points, below the 32 points of DeepSeek V4.1 Flash and GLM-5.3. API pricing is $0.70 per mill
Google DeepMind has released Gemini 4 Carbon, which ranks first on the Artificial Analysis Intelligence Index v5 with a score of 59 and leads in 4 of 9 evaluations including HLE-Diamond and TB-Science. The model focuses on cybersecurity and long-horizon tasks, claiming 88.2% on SEC-bench Pro, 54.8% on ExploitGym, and 71.0% on CWE-bench, and is initially available only to trusted cybersecurity personnel through the Fairwind Program. In terms of specifications, it supports a 1M token context and 1M maximum output, with input covering text, images, video, and speech, and output in text; Artificial Analysis lists a standard price of $4/$20 per million input/output tokens, plus a 50% launch discount.
Google DeepMind has released Gemini 4 Carbon, which ranks first on the Artificial Analysis Intelligence Index v5 with a score of 59 and leads in 4 of 9 evaluations including HLE-Diamond and TB-Science. The model focuses on cybersecurity and long-horizon tasks, claiming 88.2% on SEC-bench Pro, 54.8% on ExploitGym, and 71.0% on CWE-bench, and is initially available only to trusted cybersecurity personnel through the Fairwind Program. In terms of specifications, it supports a 1M token context and 1
Google has launched its new-generation large model Gemini 4 Carbon, focusing on cybersecurity and long-horizon complex task handling, and it ranks first with an overall score of 59 on the Artificial Analysis Intelligence Index v5.0. On WebDev Arena, the model improved by 89 Elo points over the previous generation to 1769 points, ranking fourth; internal testing shows that its success rate in outputting secure code is more than twice that of Gemini 4 Argon. Pricing is initially the same as Gemini 4 Argon: during the promotional period, input is $2 per million tokens and output is $10 per million tokens. This week it is being made available for trial to a small number of cybersecurity practitioners on a priority basis.
Google has launched its new-generation large model Gemini 4 Carbon, focusing on cybersecurity and long-horizon complex task handling, and it ranks first with an overall score of 59 on the Artificial Analysis Intelligence Index v5.0. On WebDev Arena, the model improved by 89 Elo points over the previous generation to 1769 points, ranking fourth; internal testing shows that its success rate in outputting secure code is more than twice that of Gemini 4 Argon. Pricing is initially the same as Gemini
Google has released Gemini 4 Carbon, a frontier model aimed at cybersecurity and long-horizon coding tasks, topping 15 of Google's own 17 benchmarks, but it has not yet been made public and is available only to some cybersecurity defenders. It scores 59 on Artificial Analysis' Intelligence Index, surpassing Claude Opus 5.5 and up 9 points from Argon four weeks ago; on long-input reasoning it leads OpenAI Astra and Opus by 24 points. The introductory API price is $2 per million input tokens and $10 per million output tokens, half that of Opus 5.5, doubling after the discount period to Opus's rates. Google has not announced when it will broaden access.
Google has released Gemini 4 Carbon, a frontier model aimed at cybersecurity and long-horizon coding tasks, topping 15 of Google's own 17 benchmarks, but it has not yet been made public and is available only to some cybersecurity defenders. It scores 59 on Artificial Analysis' Intelligence Index, surpassing Claude Opus 5.5 and up 9 points from Argon four weeks ago; on long-input reasoning it leads OpenAI Astra and Opus by 24 points. The introductory API price is $2 per million input tokens and $
Mistral has released the open weights for Mistral Large 4 (Le Chonk), a 1-trillion-parameter natively multimodal model that continues reinforcement learning on top of the preview version, with improved coding and defensive cybersecurity capabilities. It scores 19.7% on FrontierCode 2.0 and 48% on the Artificial Analysis Coding Agent Index, roughly on par with GLM-5.3, but uses an average of 47% fewer tokens per task; on Terminal-Bench 4.0 it improves by 13 percentage points to 41%. The model supports three reasoning modes, uses a modified MIT-based Mistral Ecosystem Open-Weight license, and is priced via API at $0.70 per million input tokens, $0.07 for cached input, and $2.10 for output.
Mistral has released the open weights for Mistral Large 4 (Le Chonk), a 1-trillion-parameter natively multimodal model that continues reinforcement learning on top of the preview version, with improved coding and defensive cybersecurity capabilities. It scores 19.7% on FrontierCode 2.0 and 48% on the Artificial Analysis Coding Agent Index, roughly on par with GLM-5.3, but uses an average of 47% fewer tokens per task; on Terminal-Bench 4.0 it improves by 13 percentage points to 41%. The model sup
NVIDIA has released Nemotron 3.5 Transcribe, an open-weight large speech model that supports 40 languages and up to 1 hour of audio per pass, and can select output formats such as verbatim, written, speaker-labeled, contextual, and translation via text instructions. The model has approximately 34B total parameters and about 4.3B activated parameters, adopts a hybrid Mamba-Transformer MoE architecture, can run in BF16 on a single H100 80GB, and the NVFP4 version at about 21GB can run on an RTX 5090. NVIDIA self-reports an average WER of 4.71% across 9 public English Open ASR test sets, the best result among open-weight models, though commercial APIs still lead; its shortcomings are that it only supports offline operation, and the Chinese far-field meeting cpCER is 34.95%.
NVIDIA has released Nemotron 3.5 Transcribe, an open-weight large speech model that supports 40 languages and up to 1 hour of audio per pass, and can select output formats such as verbatim, written, speaker-labeled, contextual, and translation via text instructions. The model has approximately 34B total parameters and about 4.3B activated parameters, adopts a hybrid Mamba-Transformer MoE architecture, can run in BF16 on a single H100 80GB, and the NVFP4 version at about 21GB can run on an RTX 50
Baseten's Base Labs has released joey-1, a 2B decision model that directly outputs a single-token label and calibrated confidence, with a P50 end-to-end latency of 20ms, about 2.5 times faster than elephant-1 (49ms), at a price of $0.01 per million tokens, one-tenth that of mammoth-1. joey-1 is post-trained from Qwen3.5-2B and shares the same data recipe and API with elephant-1 and mammoth-1, so switching requires changing only one line. The model is now available on Baseten Model APIs, with zero data retention by default, and its weights are open-sourced on Hugging Face; Model APIs supports a 16k context window, while dedicated deployments can unlock a 262k context window and support fine-tuned weights.
Baseten's Base Labs has released joey-1, a 2B decision model that directly outputs a single-token label and calibrated confidence, with a P50 end-to-end latency of 20ms, about 2.5 times faster than elephant-1 (49ms), at a price of $0.01 per million tokens, one-tenth that of mammoth-1. joey-1 is post-trained from Qwen3.5-2B and shares the same data recipe and API with elephant-1 and mammoth-1, so switching requires changing only one line. The model is now available on Baseten Model APIs, with zer
Google has released Gemini 4 Carbon, just two weeks after the release of the previous-generation frontier model Gemini 4 Argon. According to Google, Carbon surpassed Anthropic's Claude Opus 5.5 to take first place on Artificial Analysis's Intelligence Index, while Argon had previously ranked only sixth. Carbon mainly improves complex coding tasks and cybersecurity capabilities, scoring 88.2% on SEC-bench Pro, higher than Argon's 83.4% and OpenAI GPT-6 Astra's 85.4%; however, it still trails Claude Opus 5.5 on the Coding Agent Index, Terminal-Bench, and Frontiercode. Currently, Carbon is available only to selected users such as cybersecurity defense experts, and will be opened to developers, enterprises, and consumers only after passing the U.S. government's voluntary pre-review process.
Google has released Gemini 4 Carbon, just two weeks after the release of the previous-generation frontier model Gemini 4 Argon. According to Google, Carbon surpassed Anthropic's Claude Opus 5.5 to take first place on Artificial Analysis's Intelligence Index, while Argon had previously ranked only sixth. Carbon mainly improves complex coding tasks and cybersecurity capabilities, scoring 88.2% on SEC-bench Pro, higher than Argon's 83.4% and OpenAI GPT-6 Astra's 85.4%; however, it still trails Clau
Mistral has released the weights for its trillion-parameter open-source model Mistral Large 4 (ML4, nicknamed Le Chonk), which enterprises and government agencies can download and deploy themselves; the weights are already available on Hugging Face and via its API. Mistral says the model's performance is comparable to that of the open-source model GLM-5.3, but it consumes 47% fewer tokens per task, and it ranks first among open-source models on Artificial Analysis's Coding Agent Index. Pierre Stock, Mistral's Vice President of Science, said that open weights mean organizations can deploy locally in their own environments to identify, coordinate, and respond to major security incidents in real time.
Mistral has released the weights for its trillion-parameter open-source model Mistral Large 4 (ML4, nicknamed Le Chonk), which enterprises and government agencies can download and deploy themselves; the weights are already available on Hugging Face and via its API. Mistral says the model's performance is comparable to that of the open-source model GLM-5.3, but it consumes 47% fewer tokens per task, and it ranks first among open-source models on Artificial Analysis's Coding Agent Index. Pierre St
Google released Gemini 4 Carbon, which scored 59 points on the Artificial Analysis Intelligence Index v5.0 leaderboard, surpassing Claude Opus 5.5's 58 points and GPT-6 Astra's 57 points. The model achieved 88.2%, 54.8%, and 71.0% on SEC-bench Pro, ExploitGym, and CWE-bench, respectively, and Google says it can handle engineering problems in large codebases continuously for several days. The model this week is first being made available to some cybersecurity defenders, then expanded to trusted partners through the Fairwind Program, with no timeline yet for general users; initial pricing is $2 per million input tokens and $10 per million output tokens.
Google released Gemini 4 Carbon, which scored 59 points on the Artificial Analysis Intelligence Index v5.0 leaderboard, surpassing Claude Opus 5.5's 58 points and GPT-6 Astra's 57 points. The model achieved 88.2%, 54.8%, and 71.0% on SEC-bench Pro, ExploitGym, and CWE-bench, respectively, and Google says it can handle engineering problems in large codebases continuously for several days. The model this week is first being made available to some cybersecurity defenders, then expanded to trusted p
Baseten's Base Labs has released a 2B decision model, joey-1, post-trained on Qwen3.5-2B, which directly returns labels and calibrated probabilities instead of generating text, targeting high-concurrency routing, classification, and guardrail checks. The model scores 63.30 on the public Jev Decision Index 0.4 set, the highest among models under 4B, but 8.94 points lower than the 9B elephant-1. Model APIs are priced at $0.01 per million input tokens, with a P50 latency of 20 ms for an average 2000-token request; the weights are open-sourced under the MIT license on Hugging Face and can be self-hosted on a single CUDA GPU.
Baseten's Base Labs has released a 2B decision model, joey-1, post-trained on Qwen3.5-2B, which directly returns labels and calibrated probabilities instead of generating text, targeting high-concurrency routing, classification, and guardrail checks. The model scores 63.30 on the public Jev Decision Index 0.4 set, the highest among models under 4B, but 8.94 points lower than the 9B elephant-1. Model APIs are priced at $0.01 per million input tokens, with a P50 latency of 20 ms for an average 200
Google DeepMind has released Genie 4, a new-generation world model that can turn prompts into explorable, realistic scenes in real time, and for the first time the Genie model itself generates background sounds and music, while characters in the scene can also hold real-time conversations. Users move via the keyboard and use text prompts to modify content in the scene instantly, with the scene remaining coherent across multiple inputs. In the demo, a simulated Apptronik Apollo robot plays checkers, with each move planned by Gemini 4 Argon, pointing to Genie's future use in robot training.
Google DeepMind has released Genie 4, a new-generation world model that can turn prompts into explorable, realistic scenes in real time, and for the first time the Genie model itself generates background sounds and music, while characters in the scene can also hold real-time conversations. Users move via the keyboard and use text prompts to modify content in the scene instantly, with the scene remaining coherent across multiple inputs. In the demo, a simulated Apptronik Apollo robot plays checke
Google DeepMind has released the Genie 4 world model as a research preview to trusted testers; it can generate explorable worlds in real time from text and image prompts, and natively generate ambient sound, music, and character dialogue. The model introduces Gemini 3.8 Flash as a planner within the simulation loop, responsible for breaking down goals, inspecting frames, and replanning, which DeepMind calls an early step toward hybrid world models. Genie 4 is hosted only on Google's platform and will later be opened to AI Ultra users and Gemini API partners; all video and audio carry SynthID watermarks.
Google DeepMind has released the Genie 4 world model as a research preview to trusted testers; it can generate explorable worlds in real time from text and image prompts, and natively generate ambient sound, music, and character dialogue. The model introduces Gemini 3.8 Flash as a planner within the simulation loop, responsible for breaking down goals, inspecting frames, and replanning, which DeepMind calls an early step toward hybrid world models. Genie 4 is hosted only on Google's platform and
Ant Group's AntLingAGI has released Ling 3.1 Flash 1019, an updated version of Ling 3.1 Flash, scoring 44 on the Artificial Analysis Intelligence Index v4.3, up from 41 for the previous generation. The model is a reasoning model with 560B total parameters and 25B activated parameters, supports a 1M token context window, is released under the MIT license, is available as open weights on Hugging Face, and can be accessed via InclusionAI's first-party API. In terms of agentic capabilities, AA-Briefcase v1.1's Elo rose from 1397 to 1521, GDPval-AA v2.1 rose from 1617 to 1653, and AutomationBench-AA rose from 62% to 68%; Humanity's Last Exam rose from 39% to 43%, and CritPt rose from 18% to 26%. Output tokens per task increased to 145k, up from 103k for the previous generation, with a cost of $1.49 per task and pricing of $0.30 per 1M input tokens and $0.90 per 1M output tokens.
Ant Group's AntLingAGI has released Ling 3.1 Flash 1019, an updated version of Ling 3.1 Flash, scoring 44 on the Artificial Analysis Intelligence Index v4.3, up from 41 for the previous generation. The model is a reasoning model with 560B total parameters and 25B activated parameters, supports a 1M token context window, is released under the MIT license, is available as open weights on Hugging Face, and can be accessed via InclusionAI's first-party API. In terms of agentic capabilities, AA-Brief
Google previews the Genie 4 world model, which can generate realistic three-dimensional environments based on natural language descriptions and supports editing operations such as changing the weather and adding objects with follow-up prompts. Compared with the previous generation, Genie 4 is better at rendering complex details such as fluid motion, can also generate accompanying audio such as crashing waves, and makes virtual characters move naturally and be controllable through real-time interaction. Internally, the model calls Gemini Omni to ensure visual realism and Gemini 3.8 Flash to maintain inter-frame consistency. It is currently being tested with a small number of early users and will later be offered through the Google AI Ultra subscription, the Gemini API, and Google AI Studio.
Google previews the Genie 4 world model, which can generate realistic three-dimensional environments based on natural language descriptions and supports editing operations such as changing the weather and adding objects with follow-up prompts. Compared with the previous generation, Genie 4 is better at rendering complex details such as fluid motion, can also generate accompanying audio such as crashing waves, and makes virtual characters move naturally and be controllable through real-time inter
Liquid AI has released the LFM3 series of open-weight models, all branched from LFM3-1.7B-Base, covering text, vision, audio, Omni, and the d1 decision model. LFM3-1.7B-Base adopts a three-hybrid-layer architecture, is pretrained on 22T tokens, and in the mid-training stage over the final 3.5T tokens incorporates more agent trajectories and tool-calling data, with the context window extended to 256K tokens. The company says LFM3-1.7B is about 25% faster than LFM2-1.2B at an 8K context, and scores 38.42 versus 3.02 on WebArena-Verified; LFM3-VL-2B's average score across 62 vision benchmarks rises from 57.6 to 71.4. All models are available for download on Hugging Face, along with a single-GPU fine-tuning notebook.
Liquid AI has released the LFM3 series of open-weight models, all branched from LFM3-1.7B-Base, covering text, vision, audio, Omni, and the d1 decision model. LFM3-1.7B-Base adopts a three-hybrid-layer architecture, is pretrained on 22T tokens, and in the mid-training stage over the final 3.5T tokens incorporates more agent trajectories and tool-calling data, with the context window extended to 256K tokens. The company says LFM3-1.7B is about 25% faster than LFM2-1.2B at an 8K context, and score
Google DeepMind has released the world model Genie 4, which can generate interactive worlds in real time and, for the first time, natively generate ambient sound and character dialogue. The model builds on capabilities from the Gemini series; Gemini 3.8 Flash breaks down goals in a loop and checks frames to maintain causal coherence across multi-step events, and Gemini 4 Argon has also been used for joint simulation tasks with Gemini Robotics. The company says it remains an early experimental model: over long runs, object permanence and complex physical interactions still drift, all generated content carries a SynthID watermark, and it is currently available in preview to trusted testers first.
Google DeepMind has released the world model Genie 4, which can generate interactive worlds in real time and, for the first time, natively generate ambient sound and character dialogue. The model builds on capabilities from the Gemini series; Gemini 3.8 Flash breaks down goals in a loop and checks frames to maintain causal coherence across multi-step events, and Gemini 4 Argon has also been used for joint simulation tasks with Gemini Robotics. The company says it remains an early experimental mo
OpenAI has released GPT-6.1 Terra, with API pricing at $1 per million input tokens and $5 per million output tokens, down 50% and about 58% respectively from the previous generation, and half the price of GPT-6.1 Sol. OpenAI says its task completion rate on Terminal-Bench 4.0 rose from 21.5% to 50.3%, and its error rate at the highest reasoning level on difficult factual tests dropped to 4.3%. Third-party organization Artificial Analysis measured an intelligence index of 48 points, lower than Sol's 52 points and Claude Opus 5.5's 58 points, but the average cost per task fell from $1.40 to $0.40. The model is now available via API, and Plus, Pro, Enterprise, and Edu users can use it in ChatGPT Work and Codex; ordinary Chat conversations are not yet supported.
OpenAI has released GPT-6.1 Terra, with API pricing at $1 per million input tokens and $5 per million output tokens, down 50% and about 58% respectively from the previous generation, and half the price of GPT-6.1 Sol. OpenAI says its task completion rate on Terminal-Bench 4.0 rose from 21.5% to 50.3%, and its error rate at the highest reasoning level on difficult factual tests dropped to 4.3%. Third-party organization Artificial Analysis measured an intelligence index of 48 points, lower than So
OpenAI has released GPT-6.1 Terra, with API pricing at $1 per million input tokens and $5 per million output tokens, representing a 50% and approximately 58% decrease in input and output prices respectively compared to the previous generation, and only half that of GPT-6.1 Sol. The company claims its task completion rate on Terminal-Bench 4.0 increased from 21.5% to 50.3%, and it scored 60.0% on Agents' Last Exam, close to Claude Opus 5.5's 63.2%. Artificial Analysis measured an intelligence index of 48 at the highest reasoning tier, lower than GPT-6.1 Sol's 52 and Claude Opus 5.5's 58, but the average cost per task dropped from $1.40 to $0.40. Terra is now available via API, and Plus, Pro, Enterprise, and Edu users can use it in ChatGPT Work and Codex; it is not yet supported in regular Chat conversations.
OpenAI has released GPT-6.1 Terra, with API pricing at $1 per million input tokens and $5 per million output tokens, representing a 50% and approximately 58% decrease in input and output prices respectively compared to the previous generation, and only half that of GPT-6.1 Sol. The company claims its task completion rate on Terminal-Bench 4.0 increased from 21.5% to 50.3%, and it scored 60.0% on Agents' Last Exam, close to Claude Opus 5.5's 63.2%. Artificial Analysis measured an intelligence ind
