AI release odds for GPT-6, Claude 5, Grok, and Gemini

This market will resolve to "Yes" if OpenAI's GPT-5.6 model is made available to the general public by the specified date (ET). Otherwise, this market will resolve to "No." GPT-5.6 refers to a product explicitly named GPT-5.6, or a variant that is recognized as a direct successor to GPT-5.5, similar to the progression from GPT-5.1 to GPT-5.2. (e.g., GPT-5.7, GPT-5.8, etc., would qualify toward a "Yes" resolution to this market) Qualifying releases of task-specialized models (e.g., GPT-Codex/Transcribe), cost-efficiency variants (e.g., Nano/Mini), or reasoning models of the o-series family will count for this market. Products labeled as a new flagship generation GPT-6 or similar will NOT qualify. For this market to resolve to "Yes," a qualifying model must be launched and publicly accessible, including via open beta or open rolling waitlist signups. A closed beta or any form of private access will not suffice. The release must be either clearly defined and publicly announced by OpenAI as being accessible to the general public or otherwise made publicly accessible and explicitly labeled within the company’s official website. Labeling errors, placeholder text, or version names displayed on the website that do not correspond to a model that is actually accessible to the general public will not qualify. The primary resolution source for this market will be official information from OpenAI, with additional verification from a consensus of credible reporting.

On June 12, 2026, the US government directed Anthropic to suspend access to Fable 5 for foreign nationals, whether inside or outside the United States (see: https://www.anthropic.com/news/fable-mythos-access). This market will resolve to "Yes" if President Trump, the US government, or Anthropic publicly and officially announce that this directive has been rescinded, waived, modified, or otherwise lifted such that blocking public access to Claude Fable 5 for all non-US nationals is no longer directed by the US federal government by the listed date, 11:59 PM ET. Otherwise, this market will resolve to "No." A qualifying announcement must clearly indicate that guidance from the United States federal government no longer directs Anthropic to block public access to Claude Fable 5 for all foreign nationals. A partial lifting of the restriction will qualify, provided it allows renewed public access for at least some non-US nationals. Anthropic announcements that Fable 5 will be released for public access, without announcing qualifying US government guidance, will not count. Announcements must specifically relate to the model Fable 5. Qualifying announcements may include that revised US government guidance allows public access to Fable 5 for non-US nationals with additional model safety restrictions; however, if such restrictions cause Anthropic to release the model under a different name, this will not qualify. The relevant access must be public; adjusted guidance from the United States government that access to Fable 5 may be restored for select companies will not qualify. Written public statements from Donald Trump, including posts from his personal Truth Social account, will count. Videos posted on his social media accounts will also qualify. Informal announcements, statements from unnamed sources, or leaks will not qualify. The primary resolution source will be official statements from President Trump, the US government, and Anthropic; however, a consensus of credible reporting may also be used.

In April 2026, Anthropic released the AI model "Claude Mythos 5" in limited availability to approved partners in its Project Glasswing security program. On June 12, 2026, Anthropic suspended access to Mythos 5 for all customers in response to a directive from the US government. This market will resolve to "Yes" if Anthropic restores access to "Claude Mythos 5," or a model confirmed to be the same model, to at least one US partner or client whose access was previously rescinded, by the listed date, 11:59 PM ET. Otherwise, this market will resolve to "No." A qualifying model must be named "Claude Mythos 5" or be confirmed by Anthropic or by a consensus of credible reporting to be the same model. For this market to resolve to "Yes," the qualifying model (as defined above) must be restored to accessibility for at least one partner or client whose access was previously rescinded. If an otherwise qualifying restoration restricts access to certain customers based on nationality, whether geographically inside or outside of the US, that restoration will still qualify. The primary resolution source for this market will be official information from Anthropic; however, a consensus of credible reporting will also be used.

This market will resolve to “Yes” if Donald Trump signs any federal legislation into law or performs any executive action that creates, authorizes, or directs a federal government review process for the public release of new artificial intelligence models by the listed date, 11:59 PM ET. Otherwise, this market will resolve to “No.” A qualifying action must create a federal process for reviewing or approving the public release of new artificial intelligence models. A qualifying review process may apply to artificial intelligence models generally, only to models meeting specified criteria (e.g.capability, safety, cybersecurity, national-security, or other risk-based criteria), or to models selected for review at the discretion of the federal government. Legislation or executive actions which create a group or committee responsible for overseeing artificial intelligence matters will only qualify if they explicitly create a qualifying review process. Non-binding statements, proposals, unconfirmed reports, or federal review of artificial intelligence models solely for government procurement or internal government use will not qualify. The primary resolution source will be official information from the United States federal government; however, a consensus of credible reporting may also be used.

This market will resolve according to the company which owns the model which has the highest arena rank based off the Chatbot Arena LLM Leaderboard (https://lmarena.ai/) when the table under the "Leaderboard" tab is checked on December 31, 2026, 12:00 PM ET. Results from the "Rank" section on the Leaderboard tab of https://lmarena.ai/leaderboard/text with the style control off will be used to resolve this market. Models will be ordered primarily by their leaderboard rank at the market’s check time. If two or more models are tied on rank, they will be ordered by their Arena score, including any underlying, unrounded, granular values reflected in the data below the leaderboard. If a tie remains, alphabetical order of company names as listed in this market group will be used as a final tiebreaker (e.g., if the two models are tied by exact arena score, “Google” would be ranked ahead of “xAI”). This market will resolve based on the company that occupies first place under this ranking system. The resolution source for this market is the Chatbot Arena LLM Leaderboard found at https://lmarena.ai/. If this resolution source is unavailable at check time, this market will remain open until the leaderboard comes back online and resolve based on the first check after it becomes available. If it becomes permanently unavailable, this market will resolve based on another resolution source.

This market will resolve according to the date (ET) on which OpenAI's GPT-5.6 model is made available to the general public. GPT-5.6 refers to a product explicitly named GPT-5.6, or a variant that is recognized as a direct successor to GPT-5.5, similar to the progression from GPT-5.1 to GPT-5.2. (e.g., GPT-5.7, GPT-5.8, etc., would qualify toward a "Yes" resolution to this market) Qualifying releases of task-specialized models (e.g., GPT-Codex/Transcribe), cost-efficiency variants (e.g., Nano/Mini), or reasoning models of the o-series family will count for this market. Products labeled as a new flagship generation GPT-6 or similar will NOT qualify. For this market to resolve to "Yes," a qualifying model must be launched and publicly accessible, including via open beta or open rolling waitlist signups. A closed beta or any form of private access will not suffice. The release must be either clearly defined and publicly announced by OpenAI as being accessible to the general public or otherwise made publicly accessible and explicitly labeled within the company’s official website. Labeling errors, placeholder text, or version names displayed on the website that do not correspond to a model that is actually accessible to the general public will not qualify. The primary resolution source for this market will be official information from OpenAI, with additional verification from a consensus of credible reporting.

This market will resolve to "Yes" if Google's next Gemini Pro model is made available to the general public between market creation and the specified date (ET). Otherwise, this market will resolve to “No”. Any Gemini model released after market creation and labeled as "Pro" may qualify (e.g., gemini-3.2-pro, gemini-3.5-pro, or gemini-4.0-pro-preview). Gemini models labeled only as Flash, Flash-Lite, or another non-Pro variant will not qualify. Products labeled as a GA promotion of an already-existing Preview model (e.g., gemini-3.1-pro-ga) may qualify A qualifying model must be launched and publicly accessible, including via open beta or open rolling waitlist signups. A closed beta or any form of private access will not suffice. The release must be either clearly defined and publicly announced by Google as being accessible to the general public or otherwise made publicly accessible and explicitly labeled within the company's official website. Labeling errors, placeholder text, or version names displayed on the website that do not correspond to a model that is actually accessible to the general public will not qualify. The primary resolution source for this market will be official information from Google, with additional verification from a consensus of credible reporting.

This market will resolve to "Yes" if OpenAI's GPT-6 model is made available to the general public by December 31, 2025, 11:59 PM ET. Otherwise, this market will resolve to "No." For this market to resolve to "Yes," GPT-6 must be launched and publicly accessible, including via open beta or open rolling waitlist signups. A closed beta or any form of private access will not suffice. The release must be clearly defined and publicly announced by OpenAI as being accessible to the general public. GPT-6 refers to a product explicitly named GPT-6 (e.g. ChatGPT-6o would count), or one that is recognized as a successor to GPT-5, similar to the progression from GPT-3 to GPT-4. Products labeled as GPT-5.5 or similar will not count for this market's resolution. The primary resolution source for this market will be official information from OpenAI, with additional verification from a consensus of credible reporting.

This market will resolve to "Yes" if Google releases or makes available a new Gemini reasoning flagship model to the general public by the specified date (ET). Otherwise, this market will resolve to "No." Qualifying models must be positioned by Google as a next-generation, reasoning-focused flagship within the Gemini model line. For example, qualifying models include newly released or newly made generally available (GA) reasoning-focused flagship models (e.g, Gemini 3.1 GA), or any new Pro, Deep Think, or Ultra variants (e.g., Gemini 3.2 Pro, Gemini 3.4 Deep Think, Gemini 4 Ultra). Models explicitly positioned for speed, efficiency, or low-cost inference that compromise reasoning capability will NOT qualify. This includes, but is not limited to, variants such as Gemini Flash, Flash-lite, Nano, or similar lightweight or latency-optimized models, even if released under a new Gemini version number. Specialized models for non-text modalities such as video generation (e.g., Veo), image generation (e.g., Imagen, Nano Banana), music generation (e.g., Lyria), or robotics (e.g., Gemini Robotics) will NOT qualify, even if released under a qualifying Gemini version number. A qualifying model must be launched and publicly accessible, including via open beta or open rolling waitlist signups. A closed beta or any form of private access will not suffice. The release must be either clearly defined and publicly announced by Google as being accessible to the general public or otherwise made publicly accessible and explicitly labeled within the company's official website. Labeling errors, placeholder text, or version names displayed on the website that do not correspond to a model that is actually accessible to the general public will not qualify. The primary resolution source for this market will be official information from Google, with additional verification from a consensus of credible reporting.

This market will resolve to "Yes" if Grok AI officially defeats T1 in any exhibition/competitive match of League of Legends played in 2026. Otherwise, this market will resolve to "No". If no match takes place between T1 and Grok AI or no winner can be determined for any reason by December 31, 2026, 11:59PM ET, the market will resolve to "No". The primary resolution sources for this market will be official announcements from T1, X, Riot Games (if involved); however, a consensus of credible reporting may also be used.

This market will resolve to "Yes" if xAI's Grok 4.4 model is made available to the general public by the specified date (ET). Otherwise, this market will resolve to "No." Grok 4.4 refers to a product explicitly named Grok 4.4, or a variant that is recognized as a direct successor to Grok 4.3, similar to the progression from Grok 4 to Grok 4.1 (e.g., Grok 4.5, Grok 4.6, etc., would qualify toward a "Yes" resolution to this market). Qualifying releases of task-specialized models or cost-efficiency variants will count for this market. Products labeled as a new flagship generation Grok 5 or similar will NOT qualify. For this market to resolve to "Yes," a qualifying model must be launched and publicly accessible, including via open beta or open rolling waitlist signups. A closed beta or any form of private access will not suffice. The release must be either clearly defined and publicly announced by xAI as being accessible to the general public or otherwise made publicly accessible and explicitly labeled within the company’s official website. Labeling errors, placeholder text, or version names displayed on the website that do not correspond to a model that is actually accessible to the general public will not qualify. The primary resolution source for this market will be official information from xAI, with additional verification from a consensus of credible reporting.

This market will resolve to "Yes" if the next model released by OpenAI has at least the specified score on the Arena.AI Leaderboard (arena.ai/leaderboard/text) when the table under the "Leaderboard" tab is checked on the date after the day of the release, 12:00 PM ET. Otherwise, this market will resolve to "No". Results from the "Score" column under the "Text Arena | Overall" Leaderboard tab at https://lmarena.ai/leaderboard/text with style control off will be used to resolve this market. If no qualifying score for the specified model is available on the Arena.AI Leaderboard at 12:00 PM ET following the date of the release, this market will resolve based on the first subsequent instance at which such a score becomes available on the leaderboard. If no qualifying score becomes available by the end of the seventh day following the day of the model’s release, or if no qualifying model release occurs by December 31, 2026, 11:59 PM ET, this market will resolve to "No". If multiple models are released on the same calendar date or if multiple variants of the specified model appear on the Arena.AI Leaderboard at the relevant check time (e.g., base, “Thinking,” or “Instant”), the highest-scoring variant will be used for resolution. A qualifying model must be launched and publicly accessible, including via open beta or open rolling waitlist signups. A closed beta or any form of private access will not suffice. The release must be either clearly defined and publicly announced as being accessible to the general public or otherwise made publicly accessible and explicitly labeled within the company’s official website. Labeling errors, placeholder text, or version names displayed on the website that do not correspond to a model that is actually accessible to the general public will not qualify. The resolution source for this market is the Chatbot Arena LLM Leaderboard found at arena.ai/leaderboard/text. If this resolution source is unavailable at 12:00 PM ET following the date of the release, this market will resolve based on the first subsequent instance at which such a score becomes available on the leaderboard. If it remains unavailable through the end of the seventh day after a qualifying release, it will resolve to "No".

This market will resolve to "Yes" if Google DeepMind's Veo 4 model is made available to the general public by the listed date (ET). For this market to resolve to "Yes," Veo 4 must be launched and publicly accessible, including via open beta or open rolling waitlist signups. A closed beta or any form of private access will not suffice. The release must be clearly defined and publicly announced by Google as being accessible to the general public. Veo 4 refers to a product explicitly named Veo 4, Veo 4.0, or one that is recognized as a successor to Veo 3, similar to the progression from Veo 2 to Veo 3. Products labeled as Veo 3.2, Veo 3.5, or similar will not count for this market's resolution. The primary resolution source for this market will be official information from Google, with additional verification from a consensus of credible reporting.

This market will resolve to “Yes” if OpenAI makes a new frontier model available to the general public by the specified date (ET). Otherwise, this market will resolve to “No”. For this market to resolve to “Yes”, OpenAI’s new frontier model must be launched and publicly accessible, including via open beta or open rolling waitlist signups. A closed beta or any form of private access will not suffice. The release must be clearly defined and publicly announced by OpenAI as being accessible to the general public. A frontier model refers to a newly released OpenAI model that OpenAI describes as one of its most capable or next-generation, general-purpose flagship models. Qualifying new frontier models include successors to existing frontier models, such as GPT 5.2, which could succeed GPT 5.1 in the same way that GPT 5.1 succeeded GPT 5. Models focused on a specific task such as image generation or which are versions of a previous model optimized for a specific task (i.e. GPT 5.1-codex) or for cost-efficiency (i.e. GPT-5 mini) will not count. Qualifying frontier models which are separate from the OpenAI GPT series will count. A qualifying new model from OpenAI’s o-series (i.e. o1, o3) will count. The primary resolution source for this market will be official information from OpenAI, with additional verification from a consensus of credible reporting.

This market will resolve according to the calendar date (ET) on which Google's next Gemini Pro model is made available to the general public. Any Gemini model released after market creation and labeled as "Pro" may qualify (e.g., gemini-3.2-pro, gemini-3.5-pro, or gemini-4.0-pro-preview). Gemini models labeled only as Flash, Flash-Lite, or another non-Pro variant will not qualify. Products labeled as a GA promotion of an already-existing Preview model (e.g., gemini-3.1-pro-ga) may qualify. A qualifying model must be launched and publicly accessible, including via open beta or open rolling waitlist signups. A closed beta or any form of private access will not suffice. The release must be either clearly defined and publicly announced by Google as being accessible to the general public or otherwise made publicly accessible and explicitly labeled within the company's official website. Labeling errors, placeholder text, or version names displayed on the website that do not correspond to a model that is actually accessible to the general public will not qualify. The primary resolution source for this market will be official information from Google, with additional verification from a consensus of credible reporting.

This market will resolve to "Yes" if Moonshot AI’s Kimi K3 model is made available to the general public by the specified date, 11:59 PM ET. Otherwise, this market will resolve to "No." Kimi K3 refers to a product explicitly named Kimi K3 (e.g., Kimi K3.0 would count), or one that is recognized as the new flagship model or a successor to Kimi K2.5, consistent with the progression from Kimi K2 to Kimi K2.5. The release of any model within the Kimi K3 family will qualify as “Kimi K3”. Products labeled as Kimi K2.6 or similar incremental versions will not count for this market’s resolution. For this market to resolve to "Yes," Moonshot AI’s Kimi K3 must be launched and publicly accessible, including via open beta or open rolling waitlist signups. A closed beta or any form of private access will not suffice. The release must be clearly defined and publicly announced by Moonshot AI as being accessible to the general public. If a qualifying model is made publicly accessible and explicitly labeled with the relevant version name within the company’s official website, this will qualify as “publicly announced”. Labeling errors, placeholder text, or version names displayed on the website that do not correspond to a model that is actually accessible to the general public under the rules will not qualify. The primary resolution source for this market will be official information from Moonshot AI, with additional verification from a consensus of credible reporting.

This market will resolve according to the listed entity, which is the first to reach an Arena Score of 1550+ on the Chatbot Arena LLM Leaderboard (https://lmarena.ai/) by December 31, 2026, 11:59 PM ET. Results from the "Text Arena" section on the leaderboard/text tab of https://lmarena.ai/ with the style control unchecked (https://arena.ai/leaderboard/text/overall-no-style-control) will be used to resolve this market. If no company's model reaches 1550+ Arena Score by the specified time, this market will resolve to "None in 2026". If the first model to reach a 1550+ Arena Score belongs to any company other than the listed ones, this market will be resolved to "Other". The resolution source for this market is the Chatbot Arena LLM Leaderboard found at https://lmarena.ai/. If this resolution source becomes unavailable, the market will remain open until it is accessible again. If it becomes permanently unavailable, resolution will be based on another credible source.

This market will resolve to "Yes" if Tesla publicly launches a self-driving taxi service in California by June 30, 2026, 11:59 PM ET. Otherwise, this market will resolve to "No". Any taxi service available to the general public which operates without a human driver actively controlling the vehicle will count, regardless of membership or other financial restrictions. Services which are limited to Tesla employees or a limited test group without general access will not qualify. This market's resolution source will be a consensus of credible reporting.