Gemini 4 Argon vs Kimi K3

Gemini 4 Argon is Google DeepMind's frontier Gemini 4 model, built to sustain extended reasoning across long-horizon workflows in software engineering, enterprise knowledge work such as legal and finance, scientific analysis, and defensive cybersecurity. It accepts text, image, video, and document input and returns text, operating over a one million token context window. The model raises the maximum output length to one million tokens, compared with a 64K ceiling on earlier Gemini releases, so a single reasoning trajectory can run for hundreds of thousands of tokens before terminating.

On the visual side, Argon targets multimodal document comprehension, professional chart analysis, and retrieval of specific details from long recordings, reporting 91.7% on LVBench for long-video understanding and 71.6% on Chartography for chart reasoning. Google also reports 77.9% on DeepSWE v1.1, 51.3% on AutomationBench, 68% on CWE-bench v1, and first place on the Vals Index, which aggregates finance, coding, legal, and tax tasks. Following Gemini API conventions, it localizes objects by emitting bounding boxes as [ymin, xmin, ymax, xmax] integers normalized to a 0 to 1000 range, which supports prompt driven detection and grounding alongside captioning, document parsing, and visual question answering. Google describes adversarial training and automated red teaming for indirect prompt injection resistance, plus chain-of-thought and action monitors that halt execution when behavior drifts outside user intent.

Kimi K3 is a sparse Mixture-of-Experts large language model developed by Moonshot AI, with 2.8 trillion total parameters and a 1-million-token context window. The model activates 16 out of 896 experts per token using the Stable LatentMoE framework, and is built on two architectural innovations: Kimi Delta Attention (KDA), a hybrid linear attention mechanism that enables up to 6.3x faster decoding in long-context settings, and Attention Residuals (AttnRes), which selectively retrieves representations across model depth and delivers roughly 25% higher training efficiency. Together with refined training and data recipes, these structural advances yield approximately 2.5x better overall scaling efficiency compared to its predecessor Kimi K2. The model applies quantization-aware training from the supervised fine-tuning stage onward, using MXFP4 weights with MXFP8 activations for hardware compatibility. Thinking mode is always enabled at launch, with reasoning effort configurable via the reasoning_effort field.

Kimi K3 supports native visual understanding alongside text, accepting image inputs for tasks that combine software engineering and visual reasoning. It targets long-horizon coding, knowledge work, and agentic workflows, and ships in two variants: K3 Max for general chat and agent tasks, and K3 Swarm Max for large-scale parallel processing across many coordinated sub-agents. The model is compatible with the OpenAI SDK via an OpenAI-compatible API. Full model weights are scheduled for release by July 27, 2026 under a Modified MIT license, following the open-weight pattern established by the Kimi K2 model family. A technical report with full architecture, training, and evaluation details is expected to accompany the weights release.