Scores by kind of claim
Each kind of claim is scored on its own, out of 100: the share that held up, with the range it likely falls in. Fewer resolved claims give a wider range.
-
Factual claims
Statements about the present or the past, checked against the record: did they hold up?
83 77 88 -
Predictions
Statements about events outside the person's control, checked once they fell due: did they come true?
Insufficient record: 4 predictions resolved, 5 needed for a score.
-
Promises
Commitments to the person's or organisation's own future action, with something to deliver: was it delivered?
77 65 88
All types combined: 82 (likely 77 to 86, high confidence), from 94 claims: 77 factual claims, 17 promises. It mixes kinds of claim in different proportions for different people, so compare people kind by kind.
Satya Nadella's claims
51 claims
“They are. Even today, when you start saying that — the world will keep getting better in general.”
Satya Nadella, Microsoft CEO, interviewed 4 June 2026 An Interview with Microsoft CEO Satya Nadella About Finding Core Competencies (Stratechery)
Microsoft's MAI models are already fully competitive on the frontier with the other general models.
Unresolvable The public record could not settle it.
The second judge would not resolve this claim; the verdict shown is the first judge's.
Said for an organisation they lead or work for; it counts on both records.
What would count as true or false
Judge as of 2026-06-04. True: independent benchmarks or leaderboards (e.g. LMArena, Artificial Analysis) from around that date place Microsoft's MAI general models at or near the level of leading frontier models from OpenAI, Google, Anthropic and others. Mostly true: MAI models are competitive on several major benchmarks but clearly trail the top frontier models on some. Mixed: MAI models are competitive in some categories (e.g. voice or image) but not as general models. Mostly false: MAI models rank well below the frontier on general benchmarks, with only narrow strengths. False: MAI models are clearly not competitive, or no general MAI model exists.
Evidence
- Microsoft AI chief wants to eliminate Anthropic spending The Next Web (TNW) · 4 June 2026
“To back it up, Microsoft this week announced seven new in-house AI models at its annual Build conference for developers, including MAI-Thinking-1, a reasoning model it says matches Anthropic’s Claude Opus 4.6 on a widely used coding benchmark at a lower price point.”
- Microsoft AI chief wants to eliminate Anthropic spending The Next Web (TNW) · 4 June 2026
“Suleyman was candid about that. “We’ve closed an enormous gap in six months,” he told Benzinga, while acknowledging that Anthropic has released two more advanced models since Opus 4.6, giving it a lead of several months.”
- Introducing MAI-Thinking-1 | Microsoft AI Microsoft AI · 12 August 2026 Own record
“Despite this, our model is toe-to-toe with Claude Opus 4.6 on SWE-Bench Pro.”
- Best AI Image Generators in 2026: 20 Image Models Ranked - LeadWave LeadWave · 2 November 2026
“MAI-Image-2.6 (Microsoft): the best model from any company after OpenAI on LMArena's text-to-image board; Microsoft's editing preview, MAI-Image-2.7, is third for editing.”
- Best AI Image Generators in 2026: 20 Image Models Ranked - LeadWave LeadWave · 2 November 2026
“MAI-Image-2.6 (Microsoft): the best model from any company after OpenAI on LMArena's text-to-image board; Microsoft's editing preview, MAI-Image-2.7, is third for editing.”
- Best AI Image Generators in 2026: 20 Image Models Ranked - LeadWave LeadWave · 2 November 2026
“Sunburst is first on all four leaderboards, and on LMArena's editing board it leads the best model from another company, Microsoft's MAI-Image-2.7, by 46 points.”
- Best AI Image Generators in 2026: 20 Image Models Ranked - LeadWave LeadWave · 2 November 2026
“Top 10, from the leaderboard dated October 26, 2026.”
- MAI-Thinking-1: Benchmarks, Pricing & Release Date (Microsoft) | AI Model Timeline AI Model Timeline (R3ASON GmbH) · 2 June 2026
“SWE-bench Pro 52.8 (aggregator); Microsoft reported ~53% at Build 2026, on par with Claude Opus 4.6. Not yet on vals.ai for independent confirmation.”
- Artificial Analysis Intelligence Index leaderboard · benchgap benchgap · 3 November 2026
“| 147 | MAI-Thinking-1 | 18.7% | estimated ± 6.3 pp, medium confidence |”
- MAI-Thinking-1 Review 2026: Microsoft Ditches OpenAI With Its First Reasoning Model - AI Tools Hub AI Tools Hub · 2 June 2026
“Coding is competitive, not dominant. Opus 4.8 still wins SWE-Bench Pro by ~4.5 points, and Claude Code plus Dynamic Workflows still beats anything Microsoft has shipped”
- Microsoft AI model benchmark scores & progress | evals.report evals.report (inductive.ml) · 2 June 2026
“Each result is shown one benchmark at a time, with source links and evaluation dates — no blended score or composite ranking.”
Reasoning
The image-board rankings that supported "competitive in some categories" come from a leaderboard dated October 2026 (E4, E6, E7), months after the 2026-06-04 judgement date. Evidence from around that date shows only a Microsoft-reported parity with Claude Opus 4.6 on one coding benchmark, not yet independently confirmed (E1, E8), along with newer Anthropic models leading (E2, E10). That does not settle which criteria band applies.
Cite this claim
Satya Nadella, “Microsoft's MAI models are already fully competitive on the frontier with the other general models.” (said 4 June 2026). Verdict: unresolvable. Reliability API, claim c_44a47a5dcf7b, method 0.1.0+28da144b. https://epistacklabs.com/claims/c_44a47a5dcf7b
Claims without a verdict
Is a verdict here wrong, or do you speak for Satya Nadella? Reply on the claim's own page, with the form at its foot (how replies work). Verdicts are made automatically; how verdicts and scores are made.