“For example, while GPT-5.6 Sol might be the “same price” as GPT 5.5 was, it burns more than twice the amount of tokens, meaning that the “cost of intelligence” might have gone down in the sense the model is better at benchmarks, but the “cost of actually doing shit” went up.”
Our reading, the sentence that was checked:
GPT-5.6 Sol is priced like GPT-5.5 but burns more than twice the tokens, so the real cost of doing work went up.
Assessment: Unresolved Evidence as of
Where and when Ed Zitron wrote it
Where's Your Ed At, The AI Hater's Manifesto, paragraph 94.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment wyea-the-ai-haters-manifesto-594c9b-a1
We recorded 49 claims from this post. See every claim from this post.
The wording above is the version this site observed on 2026-10-28. The source may have been edited after 2026-08-25. This page does not claim the wording was captured then.
What the claim commits to
- Kind
- Observation
- Deadline, as the speaker wrote it
- None stated. This site does not add one.
- Deadline state when assessed
- No horizon
- Deadline state on 2026-10-29
- No horizon (the release’s reference date)
From the speaker
- GPT-5.6 Sol
- the “same price” as GPT 5.5 was
- more than twice the amount of tokens
- the “cost of actually doing shit” went up
Our interpretation
- The ordinary reading is that comparable completed work generally requires more than twice as many tokens and costs more with Sol.
- Equal session counts alone do not establish equal work. Compare task difficulty, completion quality, reasoning effort, harness, retries and token categories.
- The surrounding passage acknowledges price cuts. Read “same price” as referring to the earlier pricing comparison, not as denying those cuts.
- For the ongoing observation, use August 2026 through the post date as the latest reporting month. The source gives no measurement window.
- This concerns customer usage and task costs, rather than OpenAI's corporate revenue or expenditure; metric_scope is null.
Counts as supported if: Representative matched workloads show a token-use ratio above 2 and higher charges per successfully completed task, using the applicable historical prices.
Counts as contradicted if: Representative matched workloads show a ratio of 2 or less, or show that comparable completed work did not become more expensive. A different benchmark or an isolated workflow is insufficient to settle the general claim.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
Outcome: Unresolved
The record was checked, and the evidence available by the cutoff does not settle it.
Scored on a plain reading. The speaker named no measure for this claim, so it was judged on the reading an informed reader would take, stated below under our interpretation. Both reviews had to find that reading fair.
- Evidence as of
- Deadline state when assessed
- No horizon
- Assessment recorded
- Policy version
- 2026-10-28.7
Rationale
The linked article reports substantially higher tokens per session in one developer's workflow. That does not establish higher tokens per comparable completed task. Artificial Analysis found lower output-token use on its benchmark, but that measures a different workload and token category. Prices also changed before the post. The available records do not settle the broader claim about actual work costs.
Limitations
- The supporting report is a secondary account of one workflow.
- Benchmark output-token efficiency cannot directly settle total production billing.
- Restricting the claim to the cited developer's sessions would materially narrow its ordinary reading.
Evidence (3 items)
Context only: AIReiter,
a 2.32x increase
Why it matters here: The page reports 2.25 times as many tokens per session across two 14-day windows. It does not establish matched completed tasks, and the measurement dates are unspecified.
- Publisher
- AIReiter
- Kind of source
- Secondary summary
- Describes events of
- Public from
Context only: Artificial Analysis,
15k
Why it matters here: This points against treating doubled token use as a general model property. It measures benchmark output tokens, not all billed tokens in August production workflows.
- Publisher
- Artificial Analysis
- Kind of source
- Independent measurement
- Describes events of
- Public from
Context only: OpenAI,
over 20%
Why it matters here: Prices changed before publication, as Zitron also acknowledges. The earlier equal-price comparison cannot establish current task costs without recalculating the token mix and charges.
- Publisher
- OpenAI
- Kind of source
- Company disclosure: the company describing itself, not an independent finding
- Describes events of
- Public from
What would settle it
Dated matched-task records with completion quality, all billed token categories and the prices applicable to each run.
Where it was said (1)
Said once in the sources this register covers.
: Where's Your Ed At, The AI Hater's Manifesto
For example, while GPT-5.6 Sol might be the “same price” as GPT 5.5 was, it burns more than twice the amount of tokens, meaning that the “cost of intelligence” might have gone down in the sense the model is better at benchmarks, but the “cost of actually doing shit” went up.
- Where
- paragraph 94
- Published
- 2026-08-25
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment wyea-the-ai-haters-manifesto-594c9b-a1: Current. Outcome: Unresolved. Evidence as of 2026-10-28. Public date not recorded.
Recorded events
Assessment publishedBy this site
First assessment published after both reviews and the final-text check passed.