DoubleChecks

What was claimed

Your agent can now search the web 6× faster than Exa. Exa: 372ms P50. Us: 62ms P50

Our verdict

Needs caution

The 6× figure depends on a specific Exa P50 value (372 ms) that is not supported by Exa’s own published ranges (<425 ms, then <350 ms). Presenting a precise 6× speedup using an unsupported benchmark value is likely to misrepresent Exa’s performance. No source or product name is provided to verify this claim. Without knowing which agent or product is being referenced, this cannot be confirmed or denied.

All 3 AI systems agree16 sources citedChecked Jul 28, 2026

Check your own claim

Paste any statement, headline, or AI answer — 3 independent AIs verify it in seconds, with sources.

Key findings

Your agent can now search the web 6× faster than Exa based on Exa: 372ms P50 vs Us: 62ms P50.

Misleading80%
3 of 6 AIs agree·ChatGPT: Incorrect, Perplexity: Can’t verify, Perplexity: Incorrect

The agent achieves 62ms P50 latency

Can’t verify52%
All 3 AIs agree

The agent is 6× faster than Exa (372ms vs 62ms)

Verified95%
1 AI checked

Verified against 16 sources

exa.ailinkedin.comexa.aimenlotimes.comexa.aiexa.aitessl.ioproxyway.comexa.aireddit.complushcap.comfastcrw.comyoutube.comexa.aimarktechpost.comyoutube.com

Detailed Analysis

The response mixes partially correct benchmarking language with a specific numeric comparison that cannot be verified from public sources. Exa’s published P50 latency numbers do not match the 372 ms figure cited, and there is no independent data for the agent’s 62 ms P50 claim. Some claims are therefore unverified or likely inaccurate.

Why this verdict

  • The response mixes partially correct benchmarking language with a specific numeric comparison that cannot be verified from public sources.
  • Exa’s published P50 latency numbers do not match the 372 ms figure cited, and there is no independent data for the agent’s 62 ms P50 claim.
  • Some claims are therefore unverified or likely inaccurate.

Claims checked

Exa: 372ms P50.

Incorrect90%
Exa’s own materials describe Exa Fast at "p50 latency below 425ms" and later "sub350ms" rather than a specific 372 ms value.[1] [3] [6] [11] Other sources reference Exa Fast as <350ms, not 372ms P50.[2] [10] There is no authoritative source giving 372 ms as Exa’s P50 latency.

This claim may mislead readers without added context.

Your agent can now search the web 6× faster than Exa based on Exa: 372ms P50 vs Us: 62ms P50.

Misleading80%
The 6× figure depends on a specific Exa P50 value (372 ms) that is not supported by Exa’s own published ranges (<425 ms, then <350 ms).[1] [3] [6] Presenting a precise 6× speedup using an unsupported benchmark value is likely to misrepresent Exa’s performance.

This claim may mislead readers without added context.

Your agent can now search the web 6× faster than Exa.

Can’t verify60%
No public benchmark directly compares this agent’s web search latency to Exa’s, and no authoritative sources mention a 6× speed difference.[1] [2] [3] [8] [15] The claim cannot be confirmed or disproven with available data.

What other AI think

The claim contains a mathematically correct speed comparison (6x multiplier), but the Exa baseline figure of 372ms P50 is outdated or inaccurate. Current Exa products achieve faster latencies (Instant: 180-323ms, Fast: <350ms). The unnamed agent's 62ms claim is unverifiable without source information.

Why this verdict

  • The claim contains a mathematically correct speed comparison (6x multiplier), but the Exa baseline figure of 372ms P50 is outdated or inaccurate.
  • Current Exa products achieve faster latencies (Instant: 180-323ms, Fast: <350ms).
  • The unnamed agent's 62ms claim is unverifiable without source information.

Claims checked

Exa has 372ms P50 latency

Misleading85%
Current Exa products are faster: Exa Instant achieves under 180ms, and Exa Fast achieves below 350ms. The 372ms figure appears outdated or refers to an older product version.

This claim may mislead readers without added context.

The agent achieves 62ms P50 latency

Can’t verify50%
No source or product name is provided to verify this claim. Without knowing which agent or product is being referenced, this cannot be confirmed or denied.

The agent is 6× faster than Exa (372ms vs 62ms)

Verified95%
The arithmetic is correct: 372 ÷ 62 = 6. However, this comparison uses an inaccurate Exa baseline, making the overall claim misleading.
The statement mixes an unsupported numeric claim about Exa with an unverifiable self-reported latency. Exa publicly advertises sub-350ms P50 (and similar figures), which contradicts the specific 372ms P50 used. The 62ms P50 for "Us" has no authoritative public evidence I can find.

Why this verdict

  • The statement mixes an unsupported numeric claim about Exa with an unverifiable self-reported latency.
  • Exa publicly advertises sub-350ms P50 (and similar figures), which contradicts the specific 372ms P50 used.
  • The 62ms P50 for "Us" has no authoritative public evidence I can find.

Claims checked

Exa: 372ms P50.

Incorrect85%
Exa's public materials report sub-350ms (and similar <425ms) P50 latencies for their API, so the specific 372ms P50 contradicts Exa's published figures.

This claim may mislead readers without added context.

Your agent can now search the web 6× faster than Exa (372ms vs 62ms).

Misleading75%
The 6× factor is arithmetically correct if using 372/62, but it relies on a contradicted Exa figure and an unverifiable self-reported latency, so the comparative claim is misleading.

This claim may mislead readers without added context.

Us: 62ms P50.

Can’t verify45%
No independent or public measurement was found to confirm a 62ms P50 for the unnamed "Us" agent.

Share this result

Try DoubleChecks