logo

logo

Project Scope & Organization

What is Project EVA?

Project EVA (Evaluate AI) is a global Large Language Model (LLM) evaluation challenge focused on adversarial and creative testing.

Our goal: to test and push the boundaries of AI models. We invite everyone to challenge the most advanced LLMs and compete for substantial cash prizes! 💰

Updated: 2025-10-15

I've just registered. What's next?

Congratulations! You have successfully created your RemoExperts Superhuman Profile. You will soon receive an email and a platform notification with detailed information about the competition format and schedule.

Updated: 2025-10-15

Why RemoExperts? Who are the organizers?

Project EVA is initiated by 2077AI. RemoExperts (REX) hosts the registration and talent pool, serving as the sole official entry point for the event. Abaka AI provides the necessary track and platform support. By registering on REX, you create a "Superhuman Profile" to track your achievements. All subsequent challenge assignments and prize distributions will be managed and tracked on the Rex platform.

Updated: 2025-10-15

When does it start?

We are currently in the pre-registration phase. This season's launch date will be announced later, and you will be notified immediately.

Updated: 2025-10-15

What are the rewards?

The total prize pool is up to $10.24 million! This includes the Grand Annual Prize, Seasonal Championships, Creativity Awards, Community Star Awards, and more.

In addition to cash prizes, we also offer non-monetary rewards such as paper authorship, invitations to the 2077AI offline summit, and digital honor badges.

Updated: 2025-10-15

Registration & Eligibility

Who can participate?

The challenge is open globally to individuals and small teams (2–3 members). We welcome students, researchers, industry engineers, and interdisciplinary talents. Small "squads" or "guilds" of 2–3 people can also collaborate on designing cross-domain challenges.

Updated: 2025-10-15

What do I need to register?

Basic information, a resume, education details, and skill sets

  • Optional: research interests, areas of expertise, Github/Google Scholar link, and personal website

Updated: 2025-10-15

Are there any geographical or age restrictions?

No, there are no geographical restrictions.

Updated: 2025-10-15

Season Timeline

What is the current stage?

We are in the pre-registration phase, which includes open registration and expert pool enrollment. Full details on the competition format and schedule will be announced at launch.

Updated: 2025-10-15

What is the plan for Season 1 (S1)?

The theme for S1 is "The Labyrinth of Logic and Reasoning" (HLE).

  • v1.0: The season will feature a global challenge submission period, followed by a review process and a public vote, culminating in a summit and a leaderboard awards ceremony.
  • v2.0: We plan to upgrade the experience to a gamified format with points, badges, leaderboards, and even "boss battles." Stay tuned!

Updated: 2025-10-15

What do I need to submit?

Your task is to design a challenge that exposes the shortcomings of LLMs. Your submission must include:

  1. The prompt
  2. At least two examples of model failure cases
  3. Your design rationale
  4. (Optional) A challenge based on 2077AI's proprietary data

Your findings must be reproducible on at least two mainstream LLMs. Official evaluation will be conducted on models including GPT-5, Gemini 2.5 Pro, Grok-4, Seed 1.6, Claude 4, and others (subject to availability).

Updated: 2025-10-15

What are the evaluation criteria?

Submissions will be judged on:

  • Difficulty and novelty
  • Insight and depth
  • Generalizability
  • Inspirational value
  • Clarity of exposition

Updated: 2025-10-15

Awards & Prizes

What is the prize pool breakdown?

The total prize pool is $10.24M, which includes:

  • Annual Grand Champion: $500,000
  • Seasonal Prize Pool: $150,000 (for Champion, 1st Runner-up, 2nd Runner-up, Creativity Award, Methodology Award)
  • Special Contribution Award: $24,000
  • Community Stars: 50 winners × $1,000

Updated: 2025-10-15

Are there non-cash incentives?

Yes! We offer a range of valuable academic and professional rewards, including:

  • Paper authorship/acknowledgment: accumulate enough points to secure an author slot, with final ranking based on contribution
  • Invitations to the 2077AI Summit and exclusive meetups at top-tier computer science conferences
  • J-1 visiting scholar opportunities
  • Digital badges and leaderboard honors

Updated: 2025-10-15

Media & Community Collaboration

What are the official media channels and contact methods?

Updated: 2025-10-15

Want to partner with us?

You can apply to become an Official Partner, Community Ally, or Academic Pioneer. Benefits include joint exposure, shared resources, and potential funding opportunities.

Updated: 2025-10-15

How can I become a Campus or Community Ambassador?

You can apply via contact@2077ai.com. Incentives include exclusive prize opportunities, paper authorship consideration, official certification, and reward airdrops.

Updated: 2025-10-15