#105 Other

Bagel

Bagel Labs is a Toronto AI lab building distributed diffusion training infrastructure (DDM) — name and URL likely correct but the AI-research focus has no obvious Polymarket integration.

$ 574.1k 1m volume
Visit Bagel Verified Verified
Bagel homepage screenshot — Polymarket Other
Homepage www.bagel.com

Quick facts

Founded
2023
Team
1-10
Funding
seed · $8.6M raised
CoinFund (lead), Protocol Labs, Filecoin Foundation, Borderless Capital
Pricing
Unknown · $0/mo
Chains
Polygon
Venues
Polymarket
All 4 time periods

Stats across DAY · WEEK · MONTH · ALL

From the Polymarket public data API. Updated daily.

Today · 24h
#69
$2.4k
14 active users
This week · 7d
#43
$43.0k
22 active users
This month · 30d
#55
$395.2k
64 active users
All-time · all
#105
$574.1k
74 active users
On-chain attribution

builderCode

The 32-byte hash Polymarket uses to attribute volume to this builder on-chain.

0xe8358d121a1a1a849f5d81ea46d50efd006adf0ed2e7704b19c47bb7bec72b87
Profile

What is Bagel?

Bagel is the registered Polymarket Builder Program name for Bagel Labs, a Toronto-based AI research lab that builds distributed diffusion training infrastructure for generative AI workloads (image, video, world models, robotics). Bagel Labs was founded in 2023 by Bidhan Roy, has raised $8.6M with CoinFund leading and participation from Protocol Labs, Filecoin Foundation, Borderless Capital, Alumni Ventures, Future Perfect Ventures, and Breed VC, and employs approximately 9 people. Its public products are Paris-1 (the first publicly released Distributed Diffusion Model) and the Paris Inference Engine (PIE) — nothing on bagel.com references Polymarket or prediction markets. The builder code routes $574.1k across 74 all-time users with $395.2k in the past month. The product-name match plus Bagel Labs's crypto-native investor base makes this the likely identity, but the connection between Bagel Labs's AI infrastructure work and Polymarket trading volume is not publicly documented — readers should treat the linkage as identity-likely but operationally-unverified.

Bagel is the registered Polymarket builder name for Bagel Labs, a Toronto-headquartered AI research lab that pioneers Distributed Diffusion Models (DDM) — an architecture that replaces a single large diffusion model with an ensemble of smaller expert models trained independently on partitioned datasets, with no gradient synchronization between nodes. Bagel Labs was founded in 2023 by Bidhan Roy. The company has raised $8.6M (initial round led by CoinFund) with additional investors including Protocol Labs, Filecoin Foundation, Borderless Capital, Alumni Ventures, Future Perfect Ventures, and Breed VC, and reports ~9 total employees per Tracxn / Crunchbase profiles.

Known products on bagel.com are Paris-1 (the first publicly released DDM, achieving 24% FID improvement vs decentralized baselines while using 14x less data and 16x less compute), Paris-2, Paris-3, and the Paris Inference Engine (PIE). The site contains zero references to Polymarket, prediction markets, betting, or builder programs — and the team's public communications focus exclusively on generative AI research, not financial trading.

On the Polymarket builder leaderboard the bagel code is rank #55 monthly with $395.2k routed across 64 users and rank #105 all-time at $574.1k across 74 users. We could not find any blog post, X thread, or third-party coverage explaining why an AI research lab is routing six-figure prediction-market volume — the simplest explanation is that Bagel Labs runs prediction-market agents as part of their research (the firm's crypto-native investor base and Filecoin/Protocol Labs ties make this plausible) but this connection is not publicly documented.

We are publishing the URL and Bagel Labs's verified team and funding details on the basis that the name match is exact and the polymarket-uploaded logo is a JPG photo consistent with Bagel Labs branding — but flagging the operational connection between Bagel Labs's AI work and the Polymarket builder code as 'identity-likely / operationally-unverified.' If this turns out to be a separate prediction-market product that happens to share the Bagel name, the linkage should be updated.

Product

Key features

Team

Founders & team

Backing

Funding & investors

Stage
seed
Total raised
$8,600,000
Investors
CoinFund (lead), Protocol Labs, Filecoin Foundation, Borderless Capital, Alumni Ventures, Future Perfect Ventures, Breed VC, Maven11 Capital, Graph Paper Capital
Announcement
https://www.businesswire.com/news/home/20240122875776/en/Bagel-Network-a-Decentralized-ML-Data-Network-Closes-3.1M-Pre-Seed-Round-Led-by-CoinFund
Independent coverage

What others have written about Bagel

History

Rank + volume over time

Daily snapshots from our scraper. Charts populate as history accumulates.

All-time rank

Lower is better. Inverted Y-axis.

All-time volume

Cumulative routed volume in USD.

Trust signals

How to verify Bagel

  1. On-chain attribution — every Polymarket Builder routes via an order that includes their builderCode. The hash above is checkable on Polygon against Polymarket's CTF Exchange.
  2. Polymarket's own verification — $Bagel carries the official ✓ verified flag in the data API.
  3. Cross-check Polybot Arenapolybot.arena tracks wallet-level P&L for builder-tagged bots. If a builder claims specific performance, their wallet should be findable there.
  4. Tier-2 patterns to avoid — we flag builders that match the Tier-2 pattern (no public team, no verifiable claims, recurs across affiliate-SEO listicles). See our /methodology page for the full criteria.
About Bagel 08 answers

FAQ

01 What is Bagel?

Bagel is the registered Polymarket Builder Program name for Bagel Labs, a Toronto-based AI research lab that builds distributed diffusion training infrastructure for generative AI workloads (image, video, world models, robotics). Bagel Labs was founded in 2023 by Bidhan Roy, has raised $8.6M with CoinFund leading and participation from Protocol Labs, Filecoin Foundation, Borderless Capital, Alumni Ventures, Future Perfect Ventures, and Breed VC, and employs approximately 9 people. Its public products are Paris-1 (the first publicly released Distributed Diffusion Model) and the Paris Inference Engine (PIE) — nothing on bagel.com references Polymarket or prediction markets. The builder code routes $574.1k across 74 all-time users with $395.2k in the past month. The product-name match plus Bagel Labs's crypto-native investor base makes this the likely identity, but the connection between Bagel Labs's AI infrastructure work and Polymarket trading volume is not publicly documented — readers should treat the linkage as identity-likely but operationally-unverified.

02 What is the official website of Bagel?

Bagel's official site is https://www.bagel.com/. We verified the URL by visiting the homepage and confirming it references Polymarket or prediction markets directly. If you believe this URL is incorrect, email [email protected].

03 Is Bagel verified by Polymarket?

Yes — Bagel carries Polymarket's official ✓ verified flag (returned by their data API as `verified: true`). This indicates Polymarket has reviewed the builder; it's not a guarantee of product quality or financial outcomes.

04 Who founded Bagel?

Bidhan Roy (Founder). Sourced from the project's own materials and public profiles; we don't fabricate team identities and clearly mark builders whose team is publicly anonymous.

05 Has Bagel raised funding?

seed; $8.6M raised; investors include CoinFund (lead), Protocol Labs, Filecoin Foundation. Announcement: https://www.businesswire.com/news/home/20240122875776/en/Bagel-Network-a-Decentralized-ML-Data-Network-Closes-3.1M-Pre-Seed-Round-Led-by-CoinFund.

06 How much does Bagel cost?

pricing model: Unknown; paid plans from $0/month; Bagel Labs is an AI research lab, not a paid SaaS product. Polymarket connection unclear.. Sourced from the official site at the time of research; verify the current pricing on the builder's website before subscribing.

07 What are Bagel's key features?

Distributed Diffusion Models (DDM) framework; Paris-1, Paris-2, Paris-3 models; Paris Inference Engine (PIE); Heterogeneous GPU fleet compatibility; Polymarket builder code (operational link to AI work unclear). Sourced from the official site/homepage at the time of research.

08 How does Bagel compare to similar builders?

Bagel is ranked #105 all-time, with similar-category peers including Facts, whalesmarket, connex. The leaderboard reorders daily — see the history chart for Bagel's rank trajectory.