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?
80 74 86 -
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?
85 74 94 -
Guidance
A company's own figures for its results over a period, checked against what it then reported.
Insufficient record: 0 guidance claims resolved, 5 needed for a score.
All types combined: 81 (likely 76 to 86, high confidence), from 86 claims: 70 factual claims, 16 promises. It mixes kinds of claim in different proportions for different people, so compare people kind by kind.
Mark Zuckerberg's claims
40 claims
“it’s actually the latest model of a relatively smaller pre-train that we did, the internal codename Avocado”
Mark Zuckerberg, Meta CEO, interviewed on Sources podcast 15 September 2026 Transcript: Mark Zuckerberg Interview on Meta's Muse AI Agent (Sources Podcast) (The Singju Post)
Muse Spark 1.3 is the latest model from a relatively smaller pre-train Meta did, internally codenamed Avocado.
Unresolvable The public record could not settle it.
Said for an organisation they lead or work for; it counts on both records.
What would count as true or false
True if, as of 2026-09-15, Meta's latest released model in the Muse Spark line is version 1.3 and comes from a relatively smaller pre-training run internally codenamed Avocado, as shown by Meta's own announcements or credible reporting. Mostly true if the model and the Avocado codename match but a minor detail is off (for example the exact version number, or whether it is the very latest). Mostly false if Muse Spark is linked to Avocado but is from a large flagship pre-train rather than a smaller one, or the codename belongs to a different model. False if Muse Spark 1.3 does not derive from an Avocado pre-train or no such model exists.
Evidence
- Introducing Muse Spark 1.3 Meta (research.meta.ai) · 2 September 2026 Own record
“We’re excited to release Muse Spark 1.3, which delivers improved performance across agentic and coding tasks.”
- Next Muse Spark release: status and full history Next AI Model · 18 October 2026
“The newest Muse Spark model is Muse Spark 1.3, released on September 2, 2026.”
- Muse Spark 1.3: Meta Superintelligence Labs — Specs, Pricing & Review | The AI Rankings The AI Rankings · 5 August 2026
“Muse Spark 1.3 is the current version, released 3 September 2026 through Muse Code and the Meta Model API at unchanged pricing.”
- Muse Spark 1.3: Meta Superintelligence Labs — Specs, Pricing & Review | The AI Rankings The AI Rankings · 5 August 2026
“Read the release correctly: 1.2 is a coding-focused point release of 1.1, not a new base model, and it remains closed-weights with the same 1M-token context.”
- Muse Spark 1.3: Meta Superintelligence Labs — Specs, Pricing & Review | The AI Rankings The AI Rankings · 5 August 2026
“Meta shipped Muse Spark 1.3 on 3 September 2026 through Muse Code and the Meta Model API — the fourth Muse Spark release in five months.”
- Muse Spark 1.3: Meta Superintelligence Labs — Specs, Pricing & Review | The AI Rankings The AI Rankings · 5 August 2026
“On 3 September came Muse Spark 1.3, an agentic-coding point release whose strongest numbers — including the 62 Artificial Analysis measured, above GPT-6 Astra’s 61 — come from a max configuration still in limited preview.”
- Muse Spark 1.3: Meta Superintelligence Labs — Specs, Pricing & Review | The AI Rankings The AI Rankings · 5 August 2026
“Read the release correctly: 1.2 is a coding-focused point release of 1.1, not a new base model, and it remains closed-weights with the same 1M-token context.”
- Muse Spark 1.1 Evaluation Report Meta (MSL Preparedness & Red Teaming & Alignment Team, AI Security Team) · 9 July 2026 Own record
“It builds upon Muse Spark 1.0, updating the model that powers Meta AI.”
- Muse Wonder: Zuckerberg and Wang's $100B AI bidclub.ai · 12 October 2026 Own record
“Internally, Avocado was the codename for the Muse Spark line, and Watermelon was the codename for this one.”
- Muse Wonder: Zuckerberg and Wang's $100B AI bidclub.ai · 12 October 2026 Own record
“Spark 1, 1.1, 1.2 and 1.3 all share a pre-train and differ through continued RL.”
- Muse Spark 1.3 release date (September 2026): what changed and what's next Next AI Model · 2 September 2026
“Muse Spark 1.3 was released by Meta on September 2, 2026, as a flagship model in the Muse Spark line.”
- Muse Wonder: Zuckerberg and Wang's $100B AI bidclub.ai · 12 October 2026 Own record
“Wonder is the next, significantly larger pre-train, built by repeatedly climbing through model tiers roughly four or five times larger before committing to the frontier run.”
Reasoning
Independent trackers date Muse Spark 1.3's release to early September 2026 (E2, E3, E11), and E2 and E3 identify it as the newest or current Muse Spark model. Meta's own release post (E1) does not mention a pre-training run, its size, or the codename Avocado. The only sources linking the Spark line to Avocado (E9), saying the Spark versions share a pre-train (E10) and describing Wonder as a significantly larger pre-train (E12) are bidclub.ai items, which are marked as not independent of the subject; no official record or independent report in the evidence confirms these points. The central part of the claim therefore rests only on the subject's own statements.
Cite this claim
Mark Zuckerberg, “Muse Spark 1.3 is the latest model from a relatively smaller pre-train Meta did, internally codenamed Avocado.” (said 15 September 2026). Verdict: unresolvable. Reliability API, claim c_2e1dcf91db48, method 0.1.0+3f84d14d. https://epistacklabs.com/claims/c_2e1dcf91db48
Claims without a verdict
- Unresolvable 144
- Not yet due 17
- Own records only 41
- Not researched 223
- Not checkable 1,060
Is a verdict here wrong, or do you speak for Mark Zuckerberg? 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.