https://evolution.kevinbytes.com The latest additions to the LLM Tree of Life database, mapping the evolutionary tree of large language models. en-us Mon, 01 Jun 2026 00:00:00 GMT Sat, 20 Jun 2026 23:58:45 GMT https://evolution.kevinbytes.com/models/model-claude-fable5 First frontier AI model to be suspended by government order, marking a turning point in AI regulation. Demonstrated that the most capable models may face access restrictions beyond the developer's control. First model in Anthropic's 'Mythos-class' tier, representing a new capability level above Opus. Designed for long-horizon agentic work, advanced software engineering, and complex knowledge tasks. Suspended just three days after launch on June 12, 2026, after the U.S. government issued an emergency export-control directive. Mon, 01 Jun 2026 00:00:00 GMT https://evolution.kevinbytes.com/models/model-claude-fable5 anthropic-claude https://evolution.kevinbytes.com/models/model-claude4.8 Anthropic's current stable frontier model. Features a 1-million-token context window and 128K-token output — capable of processing entire codebases and generating lengthy, detailed responses. Focused improvements on honesty, more efficient tool use, and reliability for professional workflows. Fri, 01 May 2026 00:00:00 GMT https://evolution.kevinbytes.com/models/model-claude4.8 anthropic-claude https://evolution.kevinbytes.com/models/model-gemini-3.5-flash Google's speed-optimized model designed for high-volume, low-latency applications. Serves as the default model in the Gemini app, balancing strong capabilities with rapid response times and low cost per query. Fri, 01 May 2026 00:00:00 GMT https://evolution.kevinbytes.com/models/model-gemini-3.5-flash google-gemini https://evolution.kevinbytes.com/models/model-gemini-3.5-pro Google's premium reasoning flagship, unveiled at I/O 2026 with a limited rollout. Represents Google's most capable model for complex coding, analysis, and multi-step reasoning tasks. General availability expected by end of June 2026. Fri, 01 May 2026 00:00:00 GMT https://evolution.kevinbytes.com/models/model-gemini-3.5-pro google-gemini https://evolution.kevinbytes.com/models/model-grok4 xAI's current flagship model with a 1-million-token context window — enough to process roughly 750,000 words or several full-length novels at once. Features configurable reasoning effort, allowing users to choose between quick responses and deeper, more thorough analysis depending on the task. Available through both the X platform and the X.ai API. Fri, 01 May 2026 00:00:00 GMT https://evolution.kevinbytes.com/models/model-grok4 xai-grok https://evolution.kevinbytes.com/models/model-command-a-plus Largest fully open-source enterprise model at release, proving that frontier-class multimodal models could be both powerful and completely open. Cohere's most capable model, combining text and vision understanding in a sparse mixture-of-experts design (218B total parameters, 25B active). Supports 48 languages and is fully open-source under Apache 2.0, enabling deployment in government, on-premises, and air-gapped environments. Fri, 01 May 2026 00:00:00 GMT https://evolution.kevinbytes.com/models/model-command-a-plus cohere-command https://evolution.kevinbytes.com/models/model-qwen3.7 Alibaba's proprietary flagship, available only through their API. Features a massive 1 million token context window (roughly 750,000 words) and is specifically optimized for agentic workflows — tasks where the AI autonomously plans, uses tools, and completes multi-step objectives with minimal human intervention. Fri, 01 May 2026 00:00:00 GMT https://evolution.kevinbytes.com/models/model-qwen3.7 alibaba-qwen https://evolution.kevinbytes.com/models/model-nemotron3-ultra NVIDIA's answer to GPT-5 and Claude — a 550B MoE monster that uses only 55B active parameters, combining Mamba's linear-time sequence modeling with Transformer attention. NVIDIA's flagship reasoning model and the largest in the Nemotron 3 family. A 550B parameter Mixture-of-Experts model that activates only 55B parameters per computation, combining Mamba's linear-time sequence processing with Transformer attention for both efficiency and quality. Positioned as NVIDIA's answer to frontier models from OpenAI and Anthropic. Fri, 01 May 2026 00:00:00 GMT https://evolution.kevinbytes.com/models/model-nemotron3-ultra nvidia-nemotron https://evolution.kevinbytes.com/models/model-gpt5.5 OpenAI's current flagship model with a 1 million token context window. Available in Instant (speed-optimized), Thinking (deep reasoning), and Pro (maximum capability) variants. Represents OpenAI's most capable model for complex tasks, creative work, and multi-step agentic workflows. Wed, 01 Apr 2026 00:00:00 GMT https://evolution.kevinbytes.com/models/model-gpt5.5 openai-gpt https://evolution.kevinbytes.com/models/model-claude4.7 Iterative update focused on reliability — more consistent reasoning, fewer hallucinations (made-up facts), and higher-resolution image understanding. Added new cybersecurity safeguards for responsible use. Wed, 01 Apr 2026 00:00:00 GMT https://evolution.kevinbytes.com/models/model-claude4.7 anthropic-claude https://evolution.kevinbytes.com/models/model-deepseek-v4-pro DeepSeek's largest model with 1.6 trillion parameters and a 1 million token context window (roughly 750,000 words — enough to process several novels at once). Combines the architectural innovations from V3 with the reasoning capabilities of R1, representing the frontier of open-weight AI. Wed, 01 Apr 2026 00:00:00 GMT https://evolution.kevinbytes.com/models/model-deepseek-v4-pro deepseek https://evolution.kevinbytes.com/models/model-gemma4 Pushed the boundaries of what open models can do, with reasoning capabilities that rival much larger proprietary models. Edge variants enabled frontier-quality AI on mobile and embedded devices. Google's most capable open model family, featuring a hybrid architecture with configurable 'thinking' modes for reasoning tasks. Includes edge-optimized variants (E2B, E4B) small enough for mobile phones and IoT devices, alongside a 31B dense model for workstation-class performance. Released under the Apache 2.0 license for maximum flexibility. Later expanded with a 12B model capable of native audio input. Wed, 01 Apr 2026 00:00:00 GMT https://evolution.kevinbytes.com/models/model-gemma4 google-gemma https://evolution.kevinbytes.com/models/model-gpt5.3 Speed-optimized variant of the GPT-5 series designed for low-latency applications where response time matters most, such as real-time chat, autocomplete, and interactive tools. Sun, 01 Mar 2026 00:00:00 GMT https://evolution.kevinbytes.com/models/model-gpt5.3 openai-gpt https://evolution.kevinbytes.com/models/model-gpt5.4 Expanded the GPT-5 series into a full model family with Mini and Nano tiers, allowing developers to choose between maximum intelligence, balanced performance, or minimal cost depending on their application. Sun, 01 Mar 2026 00:00:00 GMT https://evolution.kevinbytes.com/models/model-gpt5.4 openai-gpt https://evolution.kevinbytes.com/models/model-mistral-small4 A compact multimodal model that can understand both text and images while excelling at reasoning and coding tasks. Open-weight and designed for efficiency — delivers frontier-level performance relative to its small size, making it ideal for cost-sensitive deployments and on-device applications. Sun, 01 Mar 2026 00:00:00 GMT https://evolution.kevinbytes.com/models/model-mistral-small4 mistral https://evolution.kevinbytes.com/models/model-mistral-medium-3.5 Mistral AI's enterprise-focused model combining advanced reasoning, code generation, and image understanding. Positioned between their smaller open models and largest offerings, it delivers strong performance on complex multi-step tasks while maintaining competitive pricing for business use cases. Sun, 01 Mar 2026 00:00:00 GMT https://evolution.kevinbytes.com/models/model-mistral-medium-3.5 mistral https://evolution.kevinbytes.com/models/model-nemotron3-super The mid-range model in NVIDIA's Nemotron 3 family, designed for multi-agent applications where multiple AI models collaborate to solve complex tasks. Uses the same hybrid Mamba-Transformer MoE architecture as Nano but scaled up to 120B total parameters with 12B active, and supports a 1 million token context window — enough to process roughly 750,000 words at once. Sun, 01 Mar 2026 00:00:00 GMT https://evolution.kevinbytes.com/models/model-nemotron3-super nvidia-nemotron https://evolution.kevinbytes.com/models/model-claude4.6 Expanded Claude's context window to 1 million tokens in beta — enough to process entire codebases or several books at once. Improved planning capabilities and reliability for long-running agent workflows. Sun, 01 Feb 2026 00:00:00 GMT https://evolution.kevinbytes.com/models/model-claude4.6 anthropic-claude https://evolution.kevinbytes.com/models/model-gemini-3.1 Pushed the context window to 2M tokens and more than doubled reasoning benchmark scores, signaling Google's commitment to both scale and intelligence improvements. Google's frontier reasoning model with roughly 1 trillion total parameters using sparse mixture-of-experts (only a fraction of the model activates per query for efficiency). Features a 2-million-token context window — enough to process several books or massive codebases simultaneously. More than doubled reasoning performance compared to Gemini 3 Pro on the ARC-AGI-2 benchmark. Sun, 01 Feb 2026 00:00:00 GMT https://evolution.kevinbytes.com/models/model-gemini-3.1 google-gemini https://evolution.kevinbytes.com/models/model-qwen3.5 An iterative refinement of Qwen 3, focused on improving code generation quality and multilingual performance. Maintained the hybrid thinking approach while delivering more reliable outputs for software development tasks and expanding language coverage. Sun, 01 Feb 2026 00:00:00 GMT https://evolution.kevinbytes.com/models/model-qwen3.5 alibaba-qwen