Microsoft AI's text-to-speech model for keeping one voice across 23 languages, with voice cloning from short audio
Agnes 3.0 Flash
New· 4Agentic coding model from Singapore-based Agnes AI, with an open-weights 33B Preview checkpoint
About Agnes 3.0 Flash
An agentic coding model from Singapore
Agnes 3.0 Flash came out on 11 September 2026 as a text model from the Singapore-based AI platform Agnes AI. Its purpose is coding done by agents and other work that runs on tools. Agnes AI's documentation puts the weight on dependable execution, tools that work as expected, memory of context across long tasks and delivery that can be trusted.
Using it through an API
The model is served under the identifier agnes-3.0-flash via Chat Completions, Responses and Messages style endpoints. It takes text and image URLs as input and can produce up to 65,536 tokens of output. A list price exists for input, output and cached input, though the documentation currently shows all three at zero with no end date given, so the list price is the figure that counts.
The context window question
On the context window of the API model, Agnes AI's own pages give two answers: 512K tokens in the documentation, and 1M tokens on the Hugging Face model card for the production model. Because the two conflict, no figure is stated here.
The open-weights Preview
Alongside it, Agnes AI put an open Preview checkpoint on Hugging Face, released under the Apache 2.0 licence. This checkpoint has 33 billion parameters and a hybrid-attention design, made up of 54 recurrent delta-rule layers and 18 global-attention layers. It handles a 262,144-token context and accepts text, image and video input. The model card is clear that the Preview is a different checkpoint from the API model. Its benchmark table, which includes 85.05 on GPQA Diamond and 74.20 on IFBench, was compiled across different harnesses and covers the Preview alone, so Agnes 3.0 Flash itself has no score here.
An outside estimate
Agnes AI cites an Intelligence Index score of 36 for the API model from Artificial Analysis, which marked the result as an estimate pending its own independent evaluation.
TrustList summary
Agnes AI, a platform based in Singapore, released Agnes 3.0 Flash on 11 September 2026. The text model is meant for agentic coding and work driven by tools. The documentation stresses stable execution, reliable tool use, keeping context over long tasks and trustworthy delivery.
What Agnes 3.0 Flash does
- API access: available under the id agnes-3.0-flash via three styles of endpoint: Chat Completions, Responses and Messages.
- Inputs and outputs: text and image-URL input, with up to 65,536 output tokens.
- Open-weights Preview: a separate 33-billion-parameter checkpoint on Hugging Face under the Apache 2.0 licence, which handles 262,144 tokens of context and can be given text, pictures or video.
- Hybrid attention design: the Preview mixes recurrent delta-rule layers with global-attention layers.
Key facts
- Maker: Agnes AI, Singapore
- Context window: Agnes AI's own pages disagree for the API model, giving 512K in the documentation and 1M on the model card
- Benchmarks: the model card's table applies to the Preview only
- Independent estimate: Artificial Analysis gave an estimated Intelligence Index score of 36, pending its own evaluation
Summary written by TrustList from the business's own published information, Nov 2026.
Benchmarks & AI stats
| Benchmark | Official | Community avg |
|---|---|---|
| No benchmark scores yet. Be the first to add one. | ||
“Official” values are editor-approved and feed the ranking. “Community avg” is the mean of member submissions (shown for transparency; it never affects the ranking until an editor approves a value).
Sign in to add a benchmark score for this model.
Agnes 3.0 Flash launches
Trust Score
An earned signal from verification, reviews, awards, transparency and engagement - the vendor can't buy it.
Updated 9/16/2026
Request a demo or quote from Agnes 3.0 Flash
Reviews
Write the first review of Agnes 3.0 Flash
Used it? Your experience helps other buyers decide.
Write a reviewQuestions & answers
No questions yet. Be the first to ask about Agnes 3.0 Flash.
Agnes 3.0 Flash alternatives
Similar listings buyers compare against this one.
Open-source 2B decision model from AWS's Strands Agents project that picks options and returns scores with a reliability figure
JetBrains' open 12B mixture-of-experts coding model with 2.5B parameters active per token under Apache 2.0
Microsoft AI's real-time speech-to-text model for voice agents, transcribing 60 languages with continuous language detection
Open-weight translation model from Bilibili's Index team, covering 150 languages and following glossary and style instructions
Liquid AI's open-weight decision model, giving yes or no, choice or score answers from text and images
Google's open multimodal embedding model that maps text, images, audio and video into one vector space, under Apache 2.0
Black Forest Labs' image generation and editing model with bounding-box layouts, multiple reference images and 4K output
More on TrustList
Everything here links back to the same verified catalogue. Pick your next stop.
- Write a reviewHelp other buyers decide
- CompaniesAgencies, consultancies and IT service providers, ranked by verified reviews.
- ProductsSoftware and SaaS with pricing, features, integrations and alternatives.
- ArticlesGuides, comparisons and research from the editorial desk and the community.
- AwardsAnnual recognition decided by verified reviews and an independent jury.
- LaunchesNew products and releases, voted up by the community every day.
- RequestsBuyers describe what they need; vendors respond directly.
- PeopleReviewers, authors and makers with public profiles.
- ComparePut up to four listings side by side before you shortlist.