$200 for ChatGPT, Claude or Gemini? I Did the Math

Value of the Information & Strength of the Argument

The video provides significant value by deconstructing a popular metric and offering a more practical comparison. It emphasizes that API-dollar equivalence is not equivalent to productivity, and introduces a task-based framework using cost per task at matched quality scores. The argumentation is solid: it uses concrete data from reputable sources, clearly explains the logic behind each calculation, and acknowledges limitations such as workload assumptions and the volatility of allowances. The inclusion of effort settings as a critical variable is an insightful contribution that most comparisons overlook. The reasoning is transparent, and the conclusions are presented with appropriate caution, making the analysis convincing and useful for consumers.

Scientific Rigor, Source Quality, Title Accuracy

The video demonstrates strong scientific rigor by relying on primary sources: SemiAnalysis for meter measurements, Artificial Analysis for task-cost benchmarks, and official pricing pages from Anthropic, OpenAI, and Google. It also references a reddit test as a weak but available data point for Google. The methodology is explained step-by-step, and the creator explicitly states the differences between API value and finished tasks. The title accurately reflects the content—the video does the math—and does not overpromise. A sponsorship segment of approximately 110 seconds (from 5:17 to 7:07) is present, but it is clearly marked and does not interfere with the analysis. The creator also highlights post-recording changes to plans, showing a commitment to accuracy.

The title accurately reflects the content: the video performs a detailed cost and value analysis of the three $200 AI plans, providing concrete math and comparisons. Slight clickbait but not misleading.

8/10

The video relies on data from SemiAnalysis, official plan pages, and Artificial Analysis, and clearly explains its methodology while acknowledging uncertainties and limitations. The reliability is high but dependent on third-party measurements that may change without notice.

• 00:00The $200 plan question
• 01:51Claude's $11,726 vs ChatGPT's $2,084
• 03:17Three $200 plans, three different 20x's
• 05:17Sponsor: TryHackMe
• 07:07How SemiAnalysis measured the meters
• 08:42API dollars vs finished tasks
• 10:39Ten $20 plans in a $200 box
• 12:06Opus medium vs Sol xhigh: 1.6x, not 5.6x
• 13:46Max effort flips the count
• 15:10Fable 5.1 vs GPT-6 Astra
• 16:43Pro 500 and Ultrafast
• 17:06Google AI Ultra: 20x, zero measurements
• 19:13The meter moves without telling you
• 20:59Who should skip the $200 tier
• 22:35Where your $200 should go

• SemiAnalysis: Anthropic Subscriptions Offer 5x+ More Value Than OpenAI — Source of the 5.6x API-value measurement and the raw allowance data used as the starting point of the analysis.
• Artificial Analysis Model Leaderboard — Provides cost-per-task estimates at matched benchmark scores, used to translate API value into finished tasks.
• Claude Pricing — Official plan details and pricing for Claude Max and Ultra.
• OpenAI: About ChatGPT Pro tiers — Official documentation on ChatGPT Pro tiers and allowances.
• Codex Pricing — Official pricing for Codex, used for API comparisons.
• API credits for Max and Ultra — Explains the $200 of API credits included with Claude Max.
• Effort defaults — Documentation on model effort settings, central to the video's analysis.
• Gemini Subscriptions — Google AI plan details.
• Gemini API Pricing — API pricing for Gemini models.
• Antigravity Models — Model availability by Google AI plan.
• Gemini Argon single-run test — User-reported measurement of Gemini AI Pro usage, used as a hint for Google Ultra value.

• Artificial Analysis Model Leaderboard — Supports the cost-per-task calculations and the finding that effort settings significantly impact value.
• Anthropic Effort Defaults — Confirms that medium is the default effort setting, validating the video's choice for comparison.