AI tool comparison
Google AI Edge Eloquent vs Sup AI
Which one should you ship with? Here is the side-by-side panel verdict, pricing read, reviewer split, and community vote comparison.
Productivity
Google AI Edge Eloquent
Free offline iOS dictation app powered by on-device Gemma ASR
75%
Panel ship
—
Community
Free
Entry
Google AI Edge Eloquent is a free iOS dictation app released quietly on April 6 with no press announcement or Product Hunt launch. It uses on-device Gemma ASR models to transcribe speech, strip filler words, and polish raw dictation into clean prose — all without an internet connection. An optional cloud mode routes cleanup through Gemini for higher quality results. Unlike competitors Wispr Flow and Willow (both $15/month), Eloquent has no subscription and no usage caps. The app is built on the same Google AI Edge framework used in Google AI Edge Gallery, suggesting it's part of a broader push to normalize on-device LLM inference on consumer hardware. The quiet launch strategy is notable: no blog post, no social announcement, just a quiet App Store submission. This kind of stealth deployment suggests Google may be seeding on-device AI use cases without the usual hype cycle — testing user retention before investing in marketing. An Android version is widely expected given the AI Edge framework's cross-platform nature.
AI Productivity
Sup AI
Runs 339 LLMs in parallel and downweights the hallucinating ones.
50%
Panel ship
—
Community
Free
Entry
Sup AI is an ensemble AI assistant that runs your query through 339 language models simultaneously, measures per-segment confidence across all responses, and synthesizes a final answer that amplifies agreement and suppresses likely hallucinations. The team claims a 52.15% score on Humanity's Last Exam (HLE) — 7.41 percentage points above the single best model — which, if verified, would make it the highest-scoring system on the benchmark to date. The underlying mechanism works like an LLM panel: each model votes on sub-claims within the response, confidence is estimated by agreement density, and the final output surfaces high-confidence segments while flagging uncertain ones. It's designed to reduce hallucination rate on factual tasks, not improve reasoning per se — the models in the ensemble aren't doing collaborative chain-of-thought, they're voting on outputs. Sup AI was built by Ken Mueller (Stanford, CEO) and Scott Mueller (AI Research Scientist) and launched on Product Hunt today. Pricing starts with $10 in free credits, no auto-charge, with a credit card required to start. The HLE benchmark claim is the headline and will face scrutiny — if verified, this is a meaningful research result. If it's cherry-picked, it's still a usable product with a differentiated architecture.
Reviewer scorecard
“The architecture here is the interesting part: Gemma ASR running fully on-device with optional cloud fallback for cleanup. This is exactly the hybrid inference pattern I'd want to build for privacy-sensitive voice apps, and Google just open-sourced the playbook by shipping it.”
“The HLE claim needs independent verification, but the underlying ensemble approach is architecturally sound for factual Q&A tasks. Running 339 models is expensive — pricing will be the gating factor for production use. The $10 free credit is a fair trial.”
“Free with no business model and no announcement sounds more like an experiment than a product. Google has a long history of quietly killing apps that don't get traction. I wouldn't build a workflow around Eloquent until it survives at least six months in the App Store.”
“Extraordinary claims require extraordinary evidence. A 7.41 point jump on HLE via ensembling — without publishing methodology — smells like benchmark gaming. The latency of running 339 models in parallel is also a real concern for anything other than async research tasks.”
“Killing the $15/month subscription model for voice AI is a meaningful shot fired. When Google ships a free, offline-first dictation app powered by on-device models, it sets a new user expectation for the whole category. Wispr and Willow are going to have to respond.”
“Model ensembling is an underexplored direction in the race to reduce hallucination. If Sup AI's approach scales, it could be more durable than fine-tuning individual models — you get the wisdom of the crowd across model families, training data, and architectures simultaneously.”
“Filler word stripping plus prose polishing in a fully offline app is genuinely useful for writers and podcasters. I dictate first drafts constantly and having this work on a plane or in a dead zone without compromising privacy is exactly what I've been waiting for.”
“For creative work, ensemble outputs tend to regress toward the mean — you get the most-agreed-upon version of something, which is usually the least interesting version. This is a tool for factual accuracy, not creativity. I'd stick with a single strong model for writing.”
Weekly AI Tool Verdicts
Get the next comparison in your inbox
New AI tools ship daily. We compare them before you waste an afternoon.