Anthropic, Claude Fable 5.2와 Mythos 5.2 출시…토큰 가격 인하
Anthropic이 새로운 Claude Fable 5.2와 제한 공개된 Mythos 5.2 모델을 공개했다. 더 빠른 출력과 토큰당 비용 절감을 제공하지만, 실제 테스트에서는 대형 프로젝트의 API 청구서가 높게 나타났다.

데이터: Polymarket, 1분마다 업데이트. · 2027년 7월 1일 마감
This market will resolve to "Yes" if the next Anthropic Claude Fable model (5.2+) added to the Arena.AI Leaderboard (https://arena.ai/leaderboard/text/overall-no-style-control) holds a rank within the specified top ranks when the table under the "Leaderboard" tab is checked at 12:00 PM ET on the calendar date following the date on which it first appears on the leaderboard. Otherwise, this market will resolve to "No".
A qualifying model must have a name or model identifier that includes "Fable" and be designated as version 5.2 or higher, regardless of capitalization or surrounding prefixes, suffixes, dates, or descriptors. For example, a version 5.2 or higher named in the same manner as Claude Fable 5 or Claude Fable 5.1 would qualify, including a new whole-number generation such as Claude Fable 6, while models whose name does not include "Fable" or which retain a version designation below 5.2, such as Claude Fable 5.1 or Claude Mythos 5.1, will not qualify.
Results from the "Rank" column under the "Text Arena | Overall" Leaderboard tab at https://arena.ai/leaderboard/text/overall-no-style-control with style control off (Adjustments: None), in the "Models" view, and with no other filters applied will be used to resolve this market. A model debuts within the specified top ranks if its displayed rank at the check time is equal to or better than the specified rank (for example, a displayed rank of 4 is within the top 5 and the top 10). This market will resolve solely based on the specified rank in the Rank column of the leaderboard, regardless of the model's score, rank spread or confidence intervals, ranks shown under other adjustments, filters, or categories, or any underlying granular or unrounded data presented elsewhere.
A model marked "AutoEval" will not be considered added to the leaderboard. A model is considered to first appear on the leaderboard on the first calendar date (ET) on which it is displayed with a rank and without the "AutoEval" label; only such ranks will be considered. An entry listed under a preliminary, code, or placeholder name is considered newly added on the first date it is displayed under a name satisfying the model specification clause above; a renamed entry's first appearance under a qualifying name counts as its addition to the leaderboard. Each separately listed entry, including reasoning or thinking-level variants of the same model, is a distinct model for the purposes of this market, and the first qualifying entry added governs resolution.
If multiple qualifying models are added to the leaderboard on the same calendar date (ET), the highest-ranking model will be used for resolution. Models added to the leaderboard on the calendar date following the initial qualifying model's first appearance will not be considered. If a qualifying model is not displayed on the leaderboard at 12:00 PM ET on the calendar date following its first appearance, that appearance will not qualify; if the model is subsequently displayed again, the date on which it reappears will be treated as a new first appearance.
A qualifying model must be newly added to the Arena.AI Leaderboard. Whether the model was previously released, publicly accessible, in beta, or otherwise available before appearing on the leaderboard is irrelevant for this market.
The resolution source for this market is the Arena.AI Leaderboard found at https://arena.ai/leaderboard/text/overall-no-style-control. If the leaderboard's URL, tab names, or control labels change, the equivalent default Text Overall view with style control disabled and no filters applied on arena.ai constitutes the same resolution source; such changes do not make the source unavailable. If this resolution source is unavailable at 12:00 PM ET on the calendar date following the date on which the qualifying model first appears on the leaderboard, this market will resolve based on the first subsequent instance at which the model's rank becomes available on the leaderboard. If it remains unavailable through the end of the seventh day after the qualifying model first appears on the leaderboard or if no qualifying model is added to the leaderboard by June 30, 2027, 11:59 PM ET, this market will resolve to "No".
판정 기준 출처: arena.ai/leaderboard/text/overall-no-style-control
Anthropic이 새로운 Claude Fable 5.2와 제한 공개된 Mythos 5.2 모델을 공개했다. 더 빠른 출력과 토큰당 비용 절감을 제공하지만, 실제 테스트에서는 대형 프로젝트의 API 청구서가 높게 나타났다.
Anthropic은 Claude 모델을 활용해 우주 MMO인 EVE Frontier에서 자율 에이전트를 운영했으며, 튜토리얼을 완료함과 동시에 팀 내 은밀한 사보타주를 테스트했습니다.
잠재적 투자자들은 기술주 매도세 속에서 Anthropic의 11월 예정 IPO에 앞서 낮은 가격과 누락된 매출 데이터를 요구하고 있다.
‘예’ 결과의 최근 가격입니다. 사건이 일어나면 1달러를 받는 계약에 트레이더가 그만큼의 센트를 지불하므로, 가격은 시장이 추정하는 확률로 읽을 수 있습니다.
시장은 아래 규칙에 따라 판정됩니다. Polymarket은 UMA 낙관적 오라클을 사용하며, 누구나 결과를 제안하고 이의 제기 기간 동안 반박할 수 있습니다.
네, Polymarket을 이용할 수 있는 지역이라면 가능합니다. 본인 지갑을 연결하면 베팅은 Polymarket으로 가고, 자금은 COINMAG를 거치지 않습니다. 베팅마다 COINMAG 수수료 1%가 추가됩니다.
Anthropic의 Claude가 이제 유료 미국 사용자에게 Plaid를 통해 은행 계좌를 연결하고 지출·잔액·청구서를 조회할 수 있게 하면서, 핀테크 창업자 Anthony Pompliano와 논쟁이 일었습니다.
Agility Robotics가 4억 6,500만 달러 규모 SPAC 합병을 통해 상장하면서 미국 최초의 휴머노이드 로봇 전문 기업이 되었다.
Riot Platforms는 텍사스 코시카나 캠퍼스에 AI 기업 Anthropic와 15년간 280억 달러 규모의 임대 계약을 체결했으며, 자금 조달이 부족할 경우 임대 크레딧에 연 12% 이자를 적용해야 하는 조항을 포함하고 있습니다.