Novig
Claude FrontierMath 25% Score: Market Locked at Certainty

Claude FrontierMath 25% Score: Market Locked at Certainty

AM Alex Mercer Crypto enthusiast
Market Resolved
Embed this market
Resolution Verdict
YES Market Resolved

Market has ended. Final implied probability: 100%.

Resolved
Volume
$76.8K
$7.6K in 24h
Liquidity
$7.8K
Low depth
7-Day Move
+0%
Stable
Time Left
Ended
Resolves Jun 30
77K Vol. Ended
25%+ $23K Vol.
100%
30%+ $12K Vol.
0%
50%+ $30K Vol.
0%
40%+ $12K Vol.
0%

Polymarket has already called this one. The contract asking whether Anthropic Claude will score at least 25% on the FrontierMath Benchmark by June 30, 2026 sits at 100% implied probability, with YES shares priced at $1.00 and NO shares effectively worthless at $0.00. That is not optimism. That is a market telling you the outcome is done.

The FrontierMath contract for the 25% threshold has traded $56,176 in total volume. Liquidity sits at $2,156. The 24-hour volume of $87 and a flat price history tell the same story: nobody is betting against this. The market has reached consensus and stopped moving.

How the FrontierMath Contract Works

This contract resolves YES if Anthropic Claude achieves a score of 25% or higher on the FrontierMath Benchmark before June 30, 2026. Polymarket serves as the resolution source. The 25% threshold is the lowest of four related contracts, which also cover 30%, 40%, and 50% targets.

  • YES: Claude scores 25%+ on FrontierMath by June 30, 2026. Price: $1.00. Probability: 100%. Resolves: June 30, 2026.
  • NO: Claude fails to reach 25% on FrontierMath by the resolution date. Price: $0.00. Probability: 0%. Resolves: June 30, 2026.

NO buyers would need Claude to miss a threshold that the market considers already cleared. The FrontierMath Benchmark, developed by EpochAI and a team of professional mathematicians, tests advanced mathematical reasoning across competition and research-level problems. Early GPT-4 class models scored below 2%. The 25% bar, once a stretch goal, is now treated as a floor. Any published score from Anthropic confirming 25% or above locks YES. A failure to publish, a methodology dispute, or a benchmark scope change are the only structural risks to NO, and the market assigns them zero weight.

Sponsored Partner
ROLRROLR

Momentum and Market Signals

Momentum here is a flat line by design. The 1-hour change is 0.0%, the 24-hour change is 0.0%, and the trend score reflects a contract that has not moved in 30 days. The 30-day high and low are both $1.00. This is what a resolved-in-all-but-name market looks like before the official close date arrives.

Total volume of $56,176 is thin by prediction market standards, and $2,156 in available liquidity means any meaningful new position would move this price on paper. But there is nowhere for this price to go. At $1.00, YES is fully priced. The $87 in 24-hour volume is residual noise, not signal. Traders are not repositioning. They are waiting for the calendar.

  • Price stability: YES has held $1.00 for 30 consecutive days. No repricing catalyst has emerged.
  • 24-hour volume ($87): Near-zero activity confirms no trader sees an edge in either direction.
  • 1-hour change (0.0%): Combined with flat 24-hour and trend data, this signals maximum conviction, not stagnation.
  • Liquidity ($2,156): Thin enough that a single large NO bet could technically push price, but no such bet has appeared.
  • Related market context: The Anthropic Claude score on Humanity’s Last Exam by June 30 contract also sits at 100%, reinforcing that the broader market treats Claude’s benchmark performance as a settled question across multiple tests.

Lines Analysis: What the FrontierMath Data Actually Shows

The case for YES is the market itself. A 100% implied probability sustained over 30 days without a single repricing event means traders who have looked at this have found nothing to challenge it. FrontierMath’s 25% threshold is the most accessible of the four tiers on Polymarket. The 40%+ contract sits at 57% implied probability and the 50%+ contract lower still, which tells you traders are calibrating carefully across the tier structure. They are not blindly bullish on Claude. They are specifically certain about 25%.

The case for NO requires believing either that Claude has not yet cleared 25% and will fail to do so before June 30, or that some resolution dispute will block the YES payout. Given that the 30%+ contract also commands significant market attention, the implied read is that Claude has already surpassed 25% in testing or published results. Open interest sits at $0, meaning no outstanding positions remain unresolved on the books. That is consistent with a contract where the informed money left long ago after pricing in the outcome.

  • FrontierMath score tiers: 30%+ at a lower probability than 25%+ confirms tiered calibration, not blanket optimism.
  • Humanity’s Last Exam contract at 100%: A second benchmark contract locked at certainty reinforces that Claude’s testing performance is broadly expected to meet multiple thresholds.
  • Open interest at $0: No live positions suggest the market has already settled in practice.
  • Volume trajectory: $56,176 total with $87 in 24 hours means the trading activity that priced this happened weeks ago.
  • Resolution date (June 30, 2026): Any official Anthropic publication or benchmark release before that date confirming 25%+ would close this immediately.

The $56,176 in total volume is modest, but it moved uniformly to one side. No counterweight emerged. That unanimity, sustained over a full month at maximum price, is the conviction signal. Thin markets can be wrong, but they are rarely wrong when every participant agrees and no new money challenges the position.

LINES VERDICT

YES: Already Priced In

The market has fully priced Claude clearing the FrontierMath 25% threshold. Every structural signal, flat price, zero NO volume, and corroborating benchmarks, points the same direction.

What the market says: Full certainty. This contract trades like a payout waiting for a calendar date, not an uncertain outcome. Volatility risk before June 30 exists only if Anthropic publishes conflicting data or the resolution source disputes methodology.

Key unknown: If Anthropic delays publishing an official FrontierMath score before June 30, 2026, the resolution source would need to adjudicate based on available evidence. A delayed or disputed publication is the one remaining risk, and the market currently prices it at zero.

This analysis reflects market conditions as of 2026-04-02. Prediction market probabilities are volatile and shift as new data and regulatory decisions emerge, especially as the June 30, 2026 resolution date approaches. Lines.com does not accept bets or provide financial or gambling advice. All market outcomes are uncertain.

Market Resolved Outcome: YES
Final Price 100%
Settled Jun 30, 2026
Duration 145 days

Resolution Analysis

YES Supporting Factors

Anthropic publishes an official FrontierMath score at or above 25% before June 30, 2026, and Polymarket resolves the contract immediately. The corroborating 100% probability on the Humanity's Last Exam contract suggests Claude's benchmark performance across multiple tests is already well-documented inside the market. Resolution would be clean and fast.

YES Risk Factors

Thin liquidity of $2,156 means a single large NO position could technically move price if new contradictory data emerged. If Anthropic delays publishing results past June 30, the resolution source faces an ambiguous call. Neither scenario is likely, but both exist as structural vulnerabilities in a low-volume contract.

NO Comeback Scenario

NO wins only if Claude fails to reach 25% on FrontierMath before the deadline, or if a benchmark methodology dispute blocks YES resolution. EpochAI revising the FrontierMath scoring criteria or Anthropic withdrawing from the benchmark program before June 30 would be the specific triggers. The market assigns this probability at zero.

Wildcard Factor

Anthropic could publish a FrontierMath score that clears 25% but falls under a contested testing protocol, triggering a Polymarket resolution dispute. Benchmark integrity challenges have affected AI evaluation markets before. If EpochAI or independent researchers flag a scoring irregularity, Polymarket's resolution panel would need to adjudicate before June 30.

Key macro factor: Rapid AI capability scaling across frontier model providers has compressed the timeline for advanced math benchmark milestones, making the 25% FrontierMath threshold significantly less ambitious in 2026 than when the benchmark launched.

Market Timeline

Jan 28, 2026
Market Created
Jan 30, 2026, 5:02 AM
Event Start
Jan 30, 2026, 5:03 AM
Market Opened
Jun 30, 2026
Market Resolution

Market Comments

Probabilities shown are market-implied and not predictions or recommendations. This content is for informational purposes only.