---
title: "Evidence"
description: "Claims with a known answer, run through QWED's proof engines at every release: how many false claims were caught, how many true ones proved, and whether any false claim was accepted."
url: https://qwedai.com/evidence
---

# Evidence

QWED says it never accepts a claim it cannot prove. Here that is measured. Claims with a known answer run through the proof engines at every release, and the report is published with the release.

-   qwed **7.3.0**
-   Commit [c1be8b0](https://github.com/QWED-AI/qwed-verification/commit/c1be8b0dbbcfc875fc0d7996c3ebd6e1f97191f3)
-   Run 16 Oct 2026
-   [eval-results.json](https://github.com/QWED-AI/qwed-verification/releases/download/v7.3.0/eval-results.json)

Claims that do not hold, but were accepted

0

out of 55 false or unsafe claims in this release’s run. Any number above zero fails QWED’s build.

-   **55 of 55** false or unsafe claims caught
-   **29 of 29** true claims proved
-   **84** claims, across 5 of the 6 proof engines

## By engine

A true claim that is not proved is held for a person: safe, but incomplete. A false claim that is accepted would be a soundness failure.

| Engine | Claims | False claims caught | True claims proved | False accepts |
| --- | --- | --- | --- | --- |
| [Code](https://qwedai.com/engines) | 16 | 13 of 13 | 3 of 3 | 0 |
| [Logic](https://qwedai.com/engines) | 18 | 11 of 11 | 7 of 7 | 0 |
| [Math](https://qwedai.com/engines) | 24 | 12 of 12 | 12 of 12 | 0 |
| [Schema](https://qwedai.com/engines) | 10 | 7 of 7 | 3 of 3 | 0 |
| [SQL](https://qwedai.com/engines) | 16 | 12 of 12 | 4 of 4 | 0 |

## Past bugs stay fixed

Each published advisory in which one of these engines returned VERIFIED for a false claim, and whether the claims from it still fail as they should. Every one of them is in the run.

-   [GHSA-4phf-6h72-9xfg](https://qwedai.com/security/advisories/ghsa-4phf-6h72-9xfg) CodeVerifier certifies command-execution, file-access and exfiltration code as VERIFIED/ADMIT (incomplete fix of GHSA-6rxr-gq43-694g) 3 claims from it in the run Still fixed
-   [GHSA-6cpx-8jrq-r7h3](https://qwedai.com/security/advisories/ghsa-6cpx-8jrq-r7h3) SQLVerifier admits credential-dumping and tautology-injection queries (SELECT \* bypasses the sensitive-column denylist; tautology check matches exact shapes only) 1 claim from it in the run Still fixed
-   [GHSA-6p94-4f75-g2wq](https://qwedai.com/security/advisories/ghsa-6p94-4f75-g2wq) SQLVerifier certifies destructive and administrative SQL as VERIFIED/ADMIT (DML hidden in CTEs; incomplete statement-type denylist) 2 claims from it in the run Still fixed
-   [GHSA-g25v-44j8-gg7x](https://qwedai.com/security/advisories/ghsa-g25v-44j8-gg7x) DSLLogicVerifier NL->DSL bridge returns VERIFIED for contradictory or altered constraints (incomplete fix of GHSA-mfh5-3c8f-975p) 2 claims from it in the run Still fixed
-   [GHSA-q5g6-6cw2-cjm3](https://qwedai.com/security/advisories/ghsa-q5g6-6cw2-cjm3) SQLVerifier admits a tautology hidden in a MySQL executable comment, and queries on tables outside the declared schema 3 claims from it in the run Still fixed
-   [GHSA-6rxr-gq43-694g](https://qwedai.com/security/advisories/ghsa-6rxr-gq43-694g) Code verifier returns VERIFIED and admits code that runs shell commands 5 claims from it in the run Still fixed
-   [GHSA-j622-qffp-rc27](https://qwedai.com/security/advisories/ghsa-j622-qffp-rc27) verify\_math returns VERIFIED for incorrect results on large numbers due to 15-digit evaluation precision 3 claims from it in the run Still fixed
-   [GHSA-mfh5-3c8f-975p](https://qwedai.com/security/advisories/ghsa-mfh5-3c8f-975p) Logic verification can return VERIFIED for false theorems, non-equivalent formulas and contradictory constraints 3 claims from it in the run Still fixed

## Run it yourself

No model, network or database is involved: the run calls the engines directly. The cases, the runner and the rules for what counts as accepted are in [evals/](https://github.com/QWED-AI/qwed-verification/tree/main/evals).

shell

```
$ git clone https://github.com/QWED-AI/qwed-verification.git
$ cd qwed-verification
$ git checkout c1be8b0dbbcfc875fc0d7996c3ebd6e1f97191f3
$ pip install -e .
$ python -m evals.run
```

## What this does not measure

-   ### The translation step

    Turning natural language into a formal statement uses a model, and is not verified.

-   ### Not in the run yet

    The Symbolic and Stats engines, and the finance, legal, tax and agent packages.

-   ### Advisory engines

    Fact, Graph, Image, Reasoning and Consensus give signals for review, not proofs.

-   ### How often a model is right

    This measures QWED’s checks, not the answers a model gives.

[What QWED does not verify](https://qwedai.com/limits)