AI tool comparison
Darwin-4B-David vs Gemini 3.1 Ultra
Which one should you ship with? Here is the side-by-side panel verdict, pricing read, reviewer split, and community vote comparison.
AI Models
Darwin-4B-David
4.5B merged model beats Gemma-4-31B on GPQA — no training needed
75%
Panel ship
—
Community
Paid
Entry
Darwin-4B-David is a 4.5-billion-parameter model that achieves 85.0% on GPQA Diamond — outperforming Google's Gemma-4-31B (84.3%) at roughly 1/7th the parameter count. The kicker: it required no training whatsoever. It was built in 45 minutes on a single H100 using MRI-guided DARE-TIES model merging, a novel variant of the merge-and-trim technique. The MRI-guided approach uses activation analysis to identify which parameters in each source model are most critical, then applies DARE-TIES merging only to the high-value weight regions. This avoids the catastrophic interference that usually degrades merged models. The result is a small model that inherits the strengths of multiple larger predecessors without any of the compute cost of fine-tuning. For the AI community, this is a meaningful data point: model merging continues to close the gap with expensive training runs. Darwin-4B-David demonstrates that thoughtful merge strategies can extract benchmark-level performance from models that are a fraction of the size, making capable AI more accessible on consumer hardware.
AI Models
Gemini 3.1 Ultra
Google's 2M-token flagship with native multimodal reasoning and sandboxed code execution
75%
Panel ship
—
Community
Paid
Entry
Gemini 3.1 Ultra is Google's most capable model to date, featuring a stable 2 million token context window — enough to process 1,500+ pages of text, hours of video, or an entire large codebase in a single session. Unlike prior Gemini versions that stitched modalities together, 3.1 Ultra was trained from the ground up to reason across text, image, audio, and video simultaneously without transcription intermediaries. It also ships with native sandboxed Python execution: write code, run it, observe the output, revise — all within a single API call. On benchmarks, Gemini 3.1 Ultra shows meaningful gains on ARC-AGI-3, GPQA Diamond, and SWE-Bench Pro, while its long-horizon planning and agentic capabilities are improved over 3.0. The 2M context window is particularly significant for enterprise use cases involving large document sets, video analysis, and extended software projects. Multimodal inputs include chart reading, diagram interpretation, and frame-by-frame video analysis. Available through the Gemini API and Google AI Ultra subscription, Gemini 3.1 Ultra positions Google squarely against OpenAI's GPT-5.5 and Anthropic's Claude Opus 4.7 at the frontier. The sandboxed code execution removes the need for third-party Code Interpreter plugins, and the model's native multimodal design means developers can pass raw audio or video without preprocessing.
Reviewer scorecard
“45 minutes on a single H100 to beat a 31B parameter model? That's an extraordinary efficiency ratio. MRI-guided merging is a technique I'll be watching closely. If this holds up across more benchmarks, it fundamentally changes how teams should think about building capable small models.”
“The native sandboxed Python execution is a major unlock. Being able to write, run, and iterate on code within the same API call — without stitching together a Code Interpreter plugin — simplifies a lot of agentic workflows. The 2M context window makes whole-repo analysis actually practical rather than theoretically possible.”
“GPQA Diamond is one benchmark. One. Benchmark performance doesn't translate linearly to real-world task performance, especially for a merged model that hasn't been fine-tuned for instruction following or RLHF alignment. Impressive number, but I'd want to see this on coding, reasoning chains, and RAG tasks before getting excited.”
“We've seen frontier model releases every few months and the benchmark improvements are getting smaller. 'Trained natively multimodal' was also claimed for Gemini 1.5 and 2.0. The 2M context window is impressive but most applications don't need it, and the cost at that scale is non-trivial. GPT-5.5 and Claude Opus 4.7 are both serious competition.”
“Model merging is the dark horse of AI efficiency research. If MRI-guided DARE-TIES merging can reliably produce results like this, it suggests we're nowhere near the ceiling for extracting value from existing open-weight models. The future may involve less training and more intelligent composition.”
“A 2M context window that natively understands video is a qualitative leap for enterprise AI. Imagine analyzing an entire quarter of earnings calls, legal discovery sets, or a full feature film for post-production — all in one shot. The sandboxed execution loop is the building block for fully autonomous data science agents.”
“A capable model in the 4-5B range that can run on a MacBook M-series is exactly what solo creators need for on-device inference. If Darwin-4B-David's performance holds on creative tasks, it's a genuine local creative AI for people without cloud budgets.”
“Native audio and video understanding without transcription intermediaries is huge for content workflows. Passing raw video directly and getting intelligent analysis — not just captions — opens up automated editing assistants, content QA, and creative research tools that weren't practical before. Google finally has a model worth building creative tools on.”
Weekly AI Tool Verdicts
Get the next comparison in your inbox
New AI tools ship daily. We compare them before you waste an afternoon.