Each sub-question estimates the probability a specific lab holds the top AA Intelligence Index spot on the resolution date; since the options are mutually exclusive and collectively exhaustive, the weighted average reflects prior expectations that will be updated by the current leaderboard state and release-cadence evidence, with final probabilities normalized across options.
## Cross-Market Signals ### No signal found
1. [sq4 | web_search | STRONG cred 70 | UP | VERY_RECENT] As of July 2026, Anthropic's Claude Fable 5 leads the AA Intelligence Index at 60, with Claude Opus 4.8 second at 56. 2. [sq4 | web_search | MODERATE cred 55 | UP | VERY_RECENT] Alternative trackers (FelloAI, Overchat) place Claude Opus 4.8 at #1 with ~61 points, ahead of GPT-5.5. 3. [sq1 | web_search | MODERATE cred 65 | DOWN | VERY_RECENT] OpenAI's GPT-5.5 (xhigh) is ranked #3 on the current Index at 55, trailing two Anthropic models. 4. [sq2 | web_search | MODERATE cred 55 | DOWN | VERY_RECENT] Google's Gemini 3.1 Pro is reported at ~57 by one tracker, behind Anthropic and OpenAI top models. 5. [sq3 | web_search | MODERATE cred 55 | DOWN | VERY_RECENT] xAI's Grok 4.3 is reported at ~53, notably behind the frontier leaders on the Index. 6. [sq4 | web_search | MODERATE cred 55 | NEUTRAL | VERY_RECENT] Claude Fable 5 was withdrawn from non-US customers June 13, 2026 due to export-control rules but remains on the leaderboard. 7. [sq5 | web_search | MODERATE cred 55 | DOWN | VERY_RECENT] GLM-5.2 (max) is the top open-weights model at 51, meaningfully behind the closed frontier leaders. 8. [sq1 | article_search | WEAK cred 60 | NEUTRAL | RECENT] OpenAI and Anthropic are both in confidential IPO filing phases with valuations projected north of $800 billion. 9. [sq2 | article_search | WEAK cred 60 | NEUTRAL | RECENT] Google I/O in May 2026 was framed as Alphabet's chance to demonstrate AI leadership; Alphabet stock up 140% over prior year. 10. [sq5 | web_search | MODERATE cred 60 | NEUTRAL | VERY_RECENT] The current top model scores ~60, below the 65 threshold cited as a clear frontier jump. ## Cross-Market Signals ### No signal found Information gaps: - No leaderboard data dated closer to August 31, 2026 - No confirmed release timelines for next-gen models (GPT-6, Gemini 4, Grok 5) before resolution - Conflicting scores between AA official surface and third-party trackers - Base rate for how often the AA Index top spot changes labs over a 2-month window Key uncertainties: - Whether OpenAI/Google/xAI release a new frontier model before Aug 31 - Whether Claude Fable 5's export withdrawal affects its leaderboard listing - How AA methodology changes could reshuffle rankings - Margins between top models are small (~1-5 pts), so a single release could flip the lead
You are an elite superforecaster. Estimate the probability of each option for this Metaculus multiple-choice question.
## Question
Which lab will have the highest score on the Artificial Analysis Intelligence Index on August 31, 2026?
## Description / Resolution Criteria
## Description
Artificial Analysis describes its Intelligence Index as a composite of multiple evaluations; recent public results put the leading model at about 60 on the index, so a threshold of 65 would indicate a clear near-term frontier jump rather than a marginal leaderboard reshuffle: https://artificialanalysis.ai/evaluations/artificial-analysis-intelligence-index
`{"format": "bot_tournament_question", "info": {"hash_id": "b841b18f24f72f78", "sheet_id": "76"}}`
## Resolution Criteria
This question resolves as the Creator with the highest Intelligence Index score of an AI model by the Artificial Analysis [LLM Leaderboard](https://artificialanalysis.ai/leaderboards/models) when checked by Metaculus on or after August 31, 2026.
## Fine Print
Please note that this question resolves based on what is displayed at the resolution source, regardless of any changes to methodology. In case of a tie or the highest score being a lab not listed in the other options, the question resolves as Other.
## Options
- Anthropic
- OpenAI
- Z AI
- Other
## Sub-question decomposition (planner)
- (w=0.30) Will OpenAI hold the highest Artificial Analysis Intelligence Index score on August 31, 2026? — OpenAI has frequently held or contested the top spot with frontier models (o-series, GPT-5) and is a leading candidate.
- (w=0.30) Will Google (DeepMind/Gemini) hold the highest Artificial Analysis Intelligence Index score on August 31, 2026? — Google's Gemini line has been highly competitive and often leads composite benchmarks; strong contender for #1.
- (w=0.20) Will xAI (Grok) hold the highest Artificial Analysis Intelligence Index score on August 31, 2026? — xAI's Grok models have recently topped or nearly topped the AA index, making them a serious contender.
- (w=0.10) Will Anthropic hold the highest Artificial Analysis Intelligence Index score on August 31, 2026? — Anthropic's Claude leads on some evals but historically lags on the composite AA index; lower but nonzero chance.
- (w=0.10) Will some other lab (e.g., DeepSeek, Qwen, Meta) hold the top score on August 31, 2026? — Chinese and open-weight labs have been rapidly closing the gap; captures tail/Other outcomes.
## Synthesized evidence
1. [sq4 | web_search | STRONG cred 70 | UP | VERY_RECENT] As of July 2026, Anthropic's Claude Fable 5 leads the AA Intelligence Index at 60, with Claude Opus 4.8 second at 56.
2. [sq4 | web_search | MODERATE cred 55 | UP | VERY_RECENT] Alternative trackers (FelloAI, Overchat) place Claude Opus 4.8 at #1 with ~61 points, ahead of GPT-5.5.
3. [sq1 | web_search | MODERATE cred 65 | DOWN | VERY_RECENT] OpenAI's GPT-5.5 (xhigh) is ranked #3 on the current Index at 55, trailing two Anthropic models.
4. [sq2 | web_search | MODERATE cred 55 | DOWN | VERY_RECENT] Google's Gemini 3.1 Pro is reported at ~57 by one tracker, behind Anthropic and OpenAI top models.
5. [sq3 | web_search | MODERATE cred 55 | DOWN | VERY_RECENT] xAI's Grok 4.3 is reported at ~53, notably behind the frontier leaders on the Index.
6. [sq4 | web_search | MODERATE cred 55 | NEUTRAL | VERY_RECENT] Claude Fable 5 was withdrawn from non-US customers June 13, 2026 due to export-control rules but remains on the leaderboard.
7. [sq5 | web_search | MODERATE cred 55 | DOWN | VERY_RECENT] GLM-5.2 (max) is the top open-weights model at 51, meaningfully behind the closed frontier leaders.
8. [sq1 | article_search | WEAK cred 60 | NEUTRAL | RECENT] OpenAI and Anthropic are both in confidential IPO filing phases with valuations projected north of $800 billion.
9. [sq2 | article_search | WEAK cred 60 | NEUTRAL | RECENT] Google I/O in May 2026 was framed as Alphabet's chance to demonstrate AI leadership; Alphabet stock up 140% over prior year.
10. [sq5 | web_search | MODERATE cred 60 | NEUTRAL | VERY_RECENT] The current top model scores ~60, below the 65 threshold cited as a clear frontier jump.
## Cross-Market Signals
### No signal found
Information gaps:
- No leaderboard data dated closer to August 31, 2026
- No confirmed release timelines for next-gen models (GPT-6, Gemini 4, Grok 5) before resolution
- Conflicting scores between AA official surface and third-party trackers
- Base rate for how often the AA Index top spot changes labs over a 2-month window
Key uncertainties:
- Whether OpenAI/Google/xAI release a new frontier model before Aug 31
- Whether Claude Fable 5's export withdrawal affects its leaderboard listing
- How AA methodology changes could reshuffle rankings
- Margins between top models are small (~1-5 pts), so a single release could flip the lead
## Required pre-forecast walkthrough
Before giving probabilities, address these explicitly in your rationale:
(a) The time left until the question resolves.
(b) The status quo outcome — the option most likely if nothing changes from today.
(c) An outcome that would be SURPRISING but plausible.
## Calibration guidance
- **Status quo anchoring.** Good forecasters put extra weight on the status quo option because the world changes slowly most of the time.
- **Leave moderate probability on most options.** Good forecasters don't drop options to ~0% — leave ≥3-5% on every option to account for unexpected outcomes, unless you have strong specific evidence ruling them out.
- Avoid uniform-distribution laziness — most options are not equally likely.
- **Anchor on prediction markets.** If liquid market prices (Polymarket / Kalshi) for these options appear in the evidence, treat them as a strong prior and stay close to them absent specific contrary evidence.
## Output
Return ONLY valid JSON, no markdown fences:
{
"rationale": "<address (a) (b) (c) above, then explain your distribution — 4-7 sentences total>",
"probabilities": {
"Anthropic": <prob>,
"OpenAI": <prob>,
"Z AI": <prob>,
"Other": <prob>
}
}
Probabilities MUST sum to ~1.0 (renormalization is applied; close enough is fine).
{
"forecast": {
"evidence_brief": "1. [sq4 | web_search | STRONG cred 70 | UP | VERY_RECENT] As of July 2026, Anthropic\u0027s Claude Fable 5 leads the AA Intelligence Index at 60, with Claude Opus 4.8 second at 56.\n2. [sq4 | web_search | MODERATE cred 55 | UP | VERY_RECENT] Alternative trackers (FelloAI, Overchat) place Claude Opus 4.8 at #1 with ~61 points, ahead of GPT-5.5.\n3. [sq1 | web_search | MODERATE cred 65 | DOWN | VERY_RECENT] OpenAI\u0027s GPT-5.5 (xhigh) is ranked #3 on the current Index at 55, trailing two Anthropic models.\n4. [sq2 | web_search | MODERATE cred 55 | DOWN | VERY_RECENT] Google\u0027s Gemini 3.1 Pro is reported at ~57 by one tracker, behind Anthropic and OpenAI top models.\n5. [sq3 | web_search | MODERATE cred 55 | DOWN | VERY_RECENT] xAI\u0027s Grok 4.3 is reported at ~53, notably behind the frontier leaders on the Index.\n6. [sq4 | web_search | MODERATE cred 55 | NEUTRAL | VERY_RECENT] Claude Fable 5 was withdrawn from non-US customers June 13, 2026 due to export-control rules but remains on the leaderboard.\n7. [sq5 | web_search | MODERATE cred 55 | DOWN | VERY_RECENT] GLM-5.2 (max) is the top open-weights model at 51, meaningfully behind the closed frontier leaders.\n8. [sq1 | article_search | WEAK cred 60 | NEUTRAL | RECENT] OpenAI and Anthropic are both in confidential IPO filing phases with valuations projected north of $800 billion.\n9. [sq2 | article_search | WEAK cred 60 | NEUTRAL | RECENT] Google I/O in May 2026 was framed as Alphabet\u0027s chance to demonstrate AI leadership; Alphabet stock up 140% over prior year.\n10. [sq5 | web_search | MODERATE cred 60 | NEUTRAL | VERY_RECENT] The current top model scores ~60, below the 65 threshold cited as a clear frontier jump.\n\n## Cross-Market Signals\n\n### No signal found\n\nInformation gaps:\n - No leaderboard data dated closer to August 31, 2026\n - No confirmed release timelines for next-gen models (GPT-6, Gemini 4, Grok 5) before resolution\n - Conflicting scores between AA official surface and third-party trackers\n - Base rate for how often the AA Index top spot changes labs over a 2-month window\n\nKey uncertainties:\n - Whether OpenAI/Google/xAI release a new frontier model before Aug 31\n - Whether Claude Fable 5\u0027s export withdrawal affects its leaderboard listing\n - How AA methodology changes could reshuffle rankings\n - Margins between top models are small (~1-5 pts), so a single release could flip the lead",
"forecast_prompt": "You are an elite superforecaster. Estimate the probability of each option for this Metaculus multiple-choice question.\n\n## Question\nWhich lab will have the highest score on the Artificial Analysis Intelligence Index on August 31, 2026?\n\n## Description / Resolution Criteria\n## Description\nArtificial Analysis describes its Intelligence Index as a composite of multiple evaluations; recent public results put the leading model at about 60 on the index, so a threshold of 65 would indicate a clear near-term frontier jump rather than a marginal leaderboard reshuffle: https://artificialanalysis.ai/evaluations/artificial-analysis-intelligence-index\n\n`{\"format\": \"bot_tournament_question\", \"info\": {\"hash_id\": \"b841b18f24f72f78\", \"sheet_id\": \"76\"}}`\n\n## Resolution Criteria\nThis question resolves as the Creator with the highest Intelligence Index score of an AI model by the Artificial Analysis [LLM Leaderboard](https://artificialanalysis.ai/leaderboards/models) when checked by Metaculus on or after August 31, 2026.\n\n## Fine Print\nPlease note that this question resolves based on what is displayed at the resolution source, regardless of any changes to methodology. In case of a tie or the highest score being a lab not listed in the other options, the question resolves as Other.\n\n## Options\n - Anthropic\n - OpenAI\n - Z AI\n - Other\n\n## Sub-question decomposition (planner)\n- (w=0.30) Will OpenAI hold the highest Artificial Analysis Intelligence Index score on August 31, 2026? \u2014 OpenAI has frequently held or contested the top spot with frontier models (o-series, GPT-5) and is a leading candidate.\n- (w=0.30) Will Google (DeepMind/Gemini) hold the highest Artificial Analysis Intelligence Index score on August 31, 2026? \u2014 Google\u0027s Gemini line has been highly competitive and often leads composite benchmarks; strong contender for #1.\n- (w=0.20) Will xAI (Grok) hold the highest Artificial Analysis Intelligence Index score on August 31, 2026? \u2014 xAI\u0027s Grok models have recently topped or nearly topped the AA index, making them a serious contender.\n- (w=0.10) Will Anthropic hold the highest Artificial Analysis Intelligence Index score on August 31, 2026? \u2014 Anthropic\u0027s Claude leads on some evals but historically lags on the composite AA index; lower but nonzero chance.\n- (w=0.10) Will some other lab (e.g., DeepSeek, Qwen, Meta) hold the top score on August 31, 2026? \u2014 Chinese and open-weight labs have been rapidly closing the gap; captures tail/Other outcomes.\n\n## Synthesized evidence\n1. [sq4 | web_search | STRONG cred 70 | UP | VERY_RECENT] As of July 2026, Anthropic\u0027s Claude Fable 5 leads the AA Intelligence Index at 60, with Claude Opus 4.8 second at 56.\n2. [sq4 | web_search | MODERATE cred 55 | UP | VERY_RECENT] Alternative trackers (FelloAI, Overchat) place Claude Opus 4.8 at #1 with ~61 points, ahead of GPT-5.5.\n3. [sq1 | web_search | MODERATE cred 65 | DOWN | VERY_RECENT] OpenAI\u0027s GPT-5.5 (xhigh) is ranked #3 on the current Index at 55, trailing two Anthropic models.\n4. [sq2 | web_search | MODERATE cred 55 | DOWN | VERY_RECENT] Google\u0027s Gemini 3.1 Pro is reported at ~57 by one tracker, behind Anthropic and OpenAI top models.\n5. [sq3 | web_search | MODERATE cred 55 | DOWN | VERY_RECENT] xAI\u0027s Grok 4.3 is reported at ~53, notably behind the frontier leaders on the Index.\n6. [sq4 | web_search | MODERATE cred 55 | NEUTRAL | VERY_RECENT] Claude Fable 5 was withdrawn from non-US customers June 13, 2026 due to export-control rules but remains on the leaderboard.\n7. [sq5 | web_search | MODERATE cred 55 | DOWN | VERY_RECENT] GLM-5.2 (max) is the top open-weights model at 51, meaningfully behind the closed frontier leaders.\n8. [sq1 | article_search | WEAK cred 60 | NEUTRAL | RECENT] OpenAI and Anthropic are both in confidential IPO filing phases with valuations projected north of $800 billion.\n9. [sq2 | article_search | WEAK cred 60 | NEUTRAL | RECENT] Google I/O in May 2026 was framed as Alphabet\u0027s chance to demonstrate AI leadership; Alphabet stock up 140% over prior year.\n10. [sq5 | web_search | MODERATE cred 60 | NEUTRAL | VERY_RECENT] The current top model scores ~60, below the 65 threshold cited as a clear frontier jump.\n\n## Cross-Market Signals\n\n### No signal found\n\nInformation gaps:\n - No leaderboard data dated closer to August 31, 2026\n - No confirmed release timelines for next-gen models (GPT-6, Gemini 4, Grok 5) before resolution\n - Conflicting scores between AA official surface and third-party trackers\n - Base rate for how often the AA Index top spot changes labs over a 2-month window\n\nKey uncertainties:\n - Whether OpenAI/Google/xAI release a new frontier model before Aug 31\n - Whether Claude Fable 5\u0027s export withdrawal affects its leaderboard listing\n - How AA methodology changes could reshuffle rankings\n - Margins between top models are small (~1-5 pts), so a single release could flip the lead\n\n## Required pre-forecast walkthrough\n\nBefore giving probabilities, address these explicitly in your rationale:\n (a) The time left until the question resolves.\n (b) The status quo outcome \u2014 the option most likely if nothing changes from today.\n (c) An outcome that would be SURPRISING but plausible.\n\n## Calibration guidance\n\n- **Status quo anchoring.** Good forecasters put extra weight on the status quo option because the world changes slowly most of the time.\n- **Leave moderate probability on most options.** Good forecasters don\u0027t drop options to ~0% \u2014 leave \u22653-5% on every option to account for unexpected outcomes, unless you have strong specific evidence ruling them out.\n- Avoid uniform-distribution laziness \u2014 most options are not equally likely.\n- **Anchor on prediction markets.** If liquid market prices (Polymarket / Kalshi) for these options appear in the evidence, treat them as a strong prior and stay close to them absent specific contrary evidence.\n\n## Output\n\nReturn ONLY valid JSON, no markdown fences:\n\n{\n \"rationale\": \"\u003caddress (a) (b) (c) above, then explain your distribution \u2014 4-7 sentences total\u003e\",\n \"probabilities\": {\n \"Anthropic\": \u003cprob\u003e,\n \"OpenAI\": \u003cprob\u003e,\n \"Z AI\": \u003cprob\u003e,\n \"Other\": \u003cprob\u003e\n }\n}\n\nProbabilities MUST sum to ~1.0 (renormalization is applied; close enough is fine).\n",
"models": [
"opus"
],
"probabilities": {
"Anthropic": 0.33,
"OpenAI": 0.27,
"Other": 0.35,
"Z AI": 0.05
},
"rationale": "(a) About two months remain until the August 31, 2026 resolution \u2014 enough time for one or two frontier model releases that could flip a lead separated by only 1-5 points. (b) Status quo per the most recent evidence has Anthropic (Claude Fable 5, ~60) narrowly leading, with Google\u0027s Gemini 3.1 Pro (~57) and OpenAI\u0027s GPT-5.5 (~55) close behind, so Anthropic is the modal single-lab outcome. (c) A surprising-but-plausible outcome is Z AI (GLM), currently the top open-weights model at ~51, leapfrogging closed frontier labs \u2014 unlikely given the gap. Because \u0027Other\u0027 bundles both Google and xAI (two serious frontier contenders), it collectively carries substantial mass despite Anthropic\u0027s current lead. Margins are razor-thin and the top spot has historically rotated between OpenAI, Google, Anthropic, and xAI, so I keep meaningful probability on OpenAI and Other. Z AI stays low but nonzero given its persistent gap to the closed frontier.",
"top_option": "Other"
},
"plan": {
"combination_logic": "weighted_average",
"domain": "tech",
"n_sub_qs": 5,
"n_tools": 3,
"reasoning_approach": "Each sub-question estimates the probability a specific lab holds the top AA Intelligence Index spot on the resolution date; since the options are mutually exclusive and collectively exhaustive, the weighted average reflects prior expectations that will be updated by the current leaderboard state and release-cadence evidence, with final probabilities normalized across options.",
"sub_questions": [
{
"id": "sq1",
"question": "Will OpenAI hold the highest Artificial Analysis Intelligence Index score on August 31, 2026?",
"rationale": "OpenAI has frequently held or contested the top spot with frontier models (o-series, GPT-5) and is a leading candidate.",
"weight": 0.3
},
{
"id": "sq2",
"question": "Will Google (DeepMind/Gemini) hold the highest Artificial Analysis Intelligence Index score on August 31, 2026?",
"rationale": "Google\u0027s Gemini line has been highly competitive and often leads composite benchmarks; strong contender for #1.",
"weight": 0.3
},
{
"id": "sq3",
"question": "Will xAI (Grok) hold the highest Artificial Analysis Intelligence Index score on August 31, 2026?",
"rationale": "xAI\u0027s Grok models have recently topped or nearly topped the AA index, making them a serious contender.",
"weight": 0.2
},
{
"id": "sq4",
"question": "Will Anthropic hold the highest Artificial Analysis Intelligence Index score on August 31, 2026?",
"rationale": "Anthropic\u0027s Claude leads on some evals but historically lags on the composite AA index; lower but nonzero chance.",
"weight": 0.1
},
{
"id": "sq5",
"question": "Will some other lab (e.g., DeepSeek, Qwen, Meta) hold the top score on August 31, 2026?",
"rationale": "Chinese and open-weight labs have been rapidly closing the gap; captures tail/Other outcomes.",
"weight": 0.1
}
],
"tool_requests": [
{
"parameters": {
"queries": [
"Artificial Analysis Intelligence Index leaderboard top model current",
"highest Artificial Analysis Intelligence Index score 2025 2026",
"Artificial Analysis Intelligence Index leader Gemini GPT Grok"
]
},
"target_sub_questions": [
"sq1",
"sq2",
"sq3",
"sq4",
"sq5"
],
"tool_name": "web_search"
},
{
"parameters": {
"brief": "Determine which AI lab currently leads the Artificial Analysis Intelligence Index and track expected frontier model releases (OpenAI, Google Gemini, xAI Grok, Anthropic Claude, DeepSeek, Qwen) through mid-2026 that could top the index.",
"max_searches": 4,
"question_title": "Which lab will have the highest score on the Artificial Analysis Intelligence Index on August 31, 2026?"
},
"target_sub_questions": [
"sq1",
"sq2",
"sq3",
"sq4",
"sq5"
],
"tool_name": "claude_news"
},
{
"parameters": {
"lookback_days": 120,
"queries": [
"Artificial Analysis Intelligence Index new frontier model",
"OpenAI Google xAI Anthropic model release benchmark leader",
"top AI model composite benchmark 2026"
]
},
"target_sub_questions": [
"sq1",
"sq2",
"sq3",
"sq4",
"sq5"
],
"tool_name": "article_search"
}
]
},
"question": {
"close_time": "2026-07-01T09:00:00Z",
"description": "## Description\nArtificial Analysis describes its Intelligence Index as a composite of multiple evaluations; recent public results put the leading model at about 60 on the index, so a threshold of 65 would indicate a clear near-term frontier jump rather than a marginal leaderboard reshuffle: https://artificialanalysis.ai/evaluations/artificial-analysis-intelligence-index\n\n`{\"format\": \"bot_tournament_question\", \"info\": {\"hash_id\": \"b841b18f24f72f78\", \"sheet_id\": \"76\"}}`\n\n## Resolution Criteria\nThis question resolves as the Creator with the highest Intelligence Index score of an AI model by the Artificial Analysis [LLM Leaderboard](https://artificialanalysis.ai/leaderboards/models) when checked by Metaculus on or after August 31, 2026.\n\n## Fine Print\nPlease note that this question resolves based on what is displayed at the resolution source, regardless of any changes to methodology. In case of a tie or the highest score being a lab not listed in the other options, the question resolves as Other.",
"options": [
"Anthropic",
"OpenAI",
"Z AI",
"Other"
],
"question_type": "multiple_choice",
"title": "Which lab will have the highest score on the Artificial Analysis Intelligence Index on August 31, 2026?"
},
"research": {
"cross_market_brief": "## Cross-Market Signals\n\n### No signal found",
"errors": [],
"has_cross_market": true,
"n_errors": 0,
"n_tools": 3,
"tools": [
{
"elapsed_s": 58.3,
"error": null,
"success": true,
"summary": "0 results",
"tool_name": "web_search"
},
{
"elapsed_s": 40.54,
"error": null,
"success": true,
"summary": "14 citations",
"tool_name": "claude_news"
},
{
"elapsed_s": 0.78,
"error": null,
"success": true,
"summary": "27 articles",
"tool_name": "article_search"
}
]
},
"synthesis": {
"evidence": [
{
"claim": "As of July 2026, Anthropic\u0027s Claude Fable 5 leads the AA Intelligence Index at 60, with Claude Opus 4.8 second at 56.",
"credibility": 70,
"direction": "UP",
"priced_in": true,
"recency": "VERY_RECENT",
"source": "web_search",
"strength": "STRONG",
"sub_question_id": "sq4"
},
{
"claim": "Alternative trackers (FelloAI, Overchat) place Claude Opus 4.8 at #1 with ~61 points, ahead of GPT-5.5.",
"credibility": 55,
"direction": "UP",
"priced_in": true,
"recency": "VERY_RECENT",
"source": "web_search",
"strength": "MODERATE",
"sub_question_id": "sq4"
},
{
"claim": "OpenAI\u0027s GPT-5.5 (xhigh) is ranked #3 on the current Index at 55, trailing two Anthropic models.",
"credibility": 65,
"direction": "DOWN",
"priced_in": true,
"recency": "VERY_RECENT",
"source": "web_search",
"strength": "MODERATE",
"sub_question_id": "sq1"
},
{
"claim": "Google\u0027s Gemini 3.1 Pro is reported at ~57 by one tracker, behind Anthropic and OpenAI top models.",
"credibility": 55,
"direction": "DOWN",
"priced_in": true,
"recency": "VERY_RECENT",
"source": "web_search",
"strength": "MODERATE",
"sub_question_id": "sq2"
},
{
"claim": "xAI\u0027s Grok 4.3 is reported at ~53, notably behind the frontier leaders on the Index.",
"credibility": 55,
"direction": "DOWN",
"priced_in": true,
"recency": "VERY_RECENT",
"source": "web_search",
"strength": "MODERATE",
"sub_question_id": "sq3"
},
{
"claim": "Claude Fable 5 was withdrawn from non-US customers June 13, 2026 due to export-control rules but remains on the leaderboard.",
"credibility": 55,
"direction": "NEUTRAL",
"priced_in": true,
"recency": "VERY_RECENT",
"source": "web_search",
"strength": "MODERATE",
"sub_question_id": "sq4"
},
{
"claim": "GLM-5.2 (max) is the top open-weights model at 51, meaningfully behind the closed frontier leaders.",
"credibility": 55,
"direction": "DOWN",
"priced_in": true,
"recency": "VERY_RECENT",
"source": "web_search",
"strength": "MODERATE",
"sub_question_id": "sq5"
},
{
"claim": "OpenAI and Anthropic are both in confidential IPO filing phases with valuations projected north of $800 billion.",
"credibility": 60,
"direction": "NEUTRAL",
"priced_in": true,
"recency": "RECENT",
"source": "article_search",
"strength": "WEAK",
"sub_question_id": "sq1"
},
{
"claim": "Google I/O in May 2026 was framed as Alphabet\u0027s chance to demonstrate AI leadership; Alphabet stock up 140% over prior year.",
"credibility": 60,
"direction": "NEUTRAL",
"priced_in": true,
"recency": "RECENT",
"source": "article_search",
"strength": "WEAK",
"sub_question_id": "sq2"
},
{
"claim": "The current top model scores ~60, below the 65 threshold cited as a clear frontier jump.",
"credibility": 60,
"direction": "NEUTRAL",
"priced_in": true,
"recency": "VERY_RECENT",
"source": "web_search",
"strength": "MODERATE",
"sub_question_id": "sq5"
}
],
"information_gaps": [
"No leaderboard data dated closer to August 31, 2026",
"No confirmed release timelines for next-gen models (GPT-6, Gemini 4, Grok 5) before resolution",
"Conflicting scores between AA official surface and third-party trackers",
"Base rate for how often the AA Index top spot changes labs over a 2-month window"
],
"key_uncertainties": [
"Whether OpenAI/Google/xAI release a new frontier model before Aug 31",
"Whether Claude Fable 5\u0027s export withdrawal affects its leaderboard listing",
"How AA methodology changes could reshuffle rankings",
"Margins between top models are small (~1-5 pts), so a single release could flip the lead"
],
"n_evidence": 10
},
"timings": {
"forecast": 16.95,
"plan": 14.78,
"research": 58.3,
"synthesis": 18.31
}
}