Start free trial

TRUE, BUTHeadline Over Filing

A May number sold as current

Its headline 52% figure is from a May run.

“Vals AI's independent benchmarks show frontier models correctly complete fewer than 52% of real financial-analyst tasks; a16z valued the evaluator at $400M.”

Vals AI (Series A blog) / a16z, 13 August 2026

Recorded 16 August 2026. Our ruling means: literally true, materially misleading.

Read the story →

Receipts pack

The evidence file for this verdict

10 receipts: 4 support the claim, 1 contradict it, 5 add context. 2 figures on the record, 2 traced to a quoted receipt.

The receipts

Each source, the line we quote from it, its date and which way it cuts. Free to read and to cite.

  1. 1Supports the claimvals.ai ↗

    We started Vals AI to solve this problem as the independent evaluator of artificial intelligence.

    Look up an archived copy ↗

  2. 2Contradicts the claimartificiallawyer.com ↗

    customer relationship with one or more of the participants

    Look up an archived copy ↗

  3. 3Supports the claimcryptobriefing.com ↗

    $40 million in a Series A

    Look up an archived copy ↗

  4. 4Contextkucoin.com ↗

    Vals AI Finance Agent v2 Reveals GPT-5.5 Hits Just 52% Accuracy

    Look up an archived copy ↗

  5. 5Contextweb.archive.org ↗

    Muse Spark 1.2 leads Finance Agent v2 with 60.60% accuracy. Claude Opus 5 follows at 58.63%

    Look up an archived copy ↗

  6. 6Contextvals.ai ↗

    with Muse Spark Open at 61.34% and Muse Spark 1.2 at 60.60%, each at roughly $0.77 per test

    Look up an archived copy ↗

  7. 7Supports the claimvals.ai ↗

    Today we’re announcing our $40M Series A at a $400M valuation

    Look up an archived copy ↗

  8. 8Contextvals.ai ↗

    Note that Vals AI has a customer relationship with one or more of the participants.

    Look up an archived copy ↗

  9. 9Contextvals.ai ↗

    The participant vendors, Thomson Reuters, vLex, Harvey, and Vecflow, who agreed to be evaluated

    Look up an archived copy ↗

  10. 10Supports the claimvals.ai ↗

    Our results have been cited in model cards from OpenAI, Anthropic, Google, Meta, and xAI.

    Look up an archived copy ↗

Members' extras

What each receipt proves, and what would change our mind.

  • What each receipt above proves, in our words
  • The figures table, each number traced to the receipt that states it
  • The trick behind the claim, explained, with the guide to spotting it
  • What would change our verdict
  • Our call on the record, with its due date

The verdict, the opening and every receipt above stay free. Membership adds our analysis to the evidence file for every checked story.

First membership: 30 days free, then A$89 a year. One introductory trial per customer. Card required; renews annually until cancelled. Cancel before the trial ends to avoid the first charge.