{"slug": "the-context-cliff-200k-272k-and-the-line-that-doubles-your-prompt-bill", "title": "The context cliff: 200K, 272K, and the line that doubles your prompt bill", "summary": "A developer's analysis of vendor pricing pages shows that long-context tier thresholds differ across providers, creating a 72,000-token window between 200K and 272K where Google and xAI charge their doubled long-context rates but OpenAI does not. A 250K-token prompt costs $2 per 1M input on GPT-6.1 Sol but $4 on Gemini 3.1 Pro, and crossing Gemini's 200K line by two tokens raises a request's cost by roughly 96% because the whole request is re-priced. The writeup also notes that output multipliers and cache rates vary by provider, and that the published rates were checked between 2026-11-15 and 2026-11-18.", "body_md": "Most LLM price tables are read as one number per model. That number is a floor, and the ceiling is printed next to it in smaller type.\n\nTake a 250,000-token prompt. On GPT-6.1 Sol it bills at the standard $2 per 1M input. On Gemini 3.1 Pro the same prompt is over the 200K line, so input doubles to $4. The model hasn't started writing yet and the input side already costs twice as much.\n\nThe tier lines don't line up:\n\n| Model | Standard input / output | Over the line | Line | \n|---|---|---|---|\n| Claude Haiku 5.5 | $0.10 / $0.50 | $0.50 / $2.50 | 100K | \n| Gemini 3.1 Pro | $2 / $12 | $4 / $18 | 200K | \n| Grok 4.3 | $1.25 / $2.50 | $2.50 / $5 | 200K | \n| Grok 4.7 | $2 / $6 | $4 / $12 | 200K | \n| GPT-6.1 Sol | $2 / $10 | $4 / $15 | 272K | \n| GPT-6 Astra | $10 / $50 | $20 / $75 | 272K | \n\nThat leaves a 72,000-token window, from 200K to 272K, where Google and xAI charge their long-context rate and OpenAI doesn't.\n\nRun the same request at 199,999 input tokens and 200,001, with 10K output tokens:\n\nAt the lower count Gemini costs about 4% more. Two tokens later it costs about 96% more, because the whole request moved to the higher rate. Nothing about the workload changed; the prompt crossed a line that only appears in a footnote.\n\nThe output side moves too, and not by the same factor. OpenAI's long-context rows are 2x input and 1.5x output. xAI is 2x on both. Haiku 5.5 is 5x on both, and its line is at 100K, which most teams never look at because the model is the cheap one.\n\nA blended cost per million tokens is useful for a slide and useless for a budget. Suppose a workload makes 1,000 calls a month at 250K input and 10K output, all of them between the 200K and 272K lines:\n\n| Model | Per call | 1,000 calls | \n|---|---|---|\n| GPT-6.1 Sol | $0.60 | $600 | \n| Grok 4.3 | $0.68 | $675 | \n| Gemini 3.1 Pro | $1.18 | $1,180 | \n| Claude Haiku 5.5 (over its 100K line) | $0.15 | $150 | \n\nTwo calls in five landing above a threshold is enough to move the monthly number by hundreds of dollars, and the average token price you computed last quarter won't show it.\n\nCached input makes the gap harder to read. Grok's cached rate goes from $0.50 to $1.00 per 1M at the 200K line. The long-context row for GPT-6 Astra lists $20/$75, a two-number row; if most of your input is cache hits, that row alone doesn't tell you the bill, and the vendor pricing page is the only place to check the cached rate at that tier.\n\nOne caveat: whether a provider re-prices the whole request or only the tokens above the line is in the API docs, and it has changed before. The table here records the published rate for requests over the threshold. Test one long call against your own usage before building a budget on the multiplication.\n\nThe numbers above come from the vendor pricing pages, checked between 2026-11-15 and 2026-11-18. They move: Gemini 4 Carbon is already scheduled to go from $2/$10 to $4/$20 on 2026-12-01.\n\nThe full table, with a checked date on every row: [AI API price comparison](https://aiapicostcalc.com/models). If you want to price one workload instead of reading rows, the [token cost calculator](https://aiapicostcalc.com/token-cost-calculator) takes input, output, and cache-hit counts. The two models used in the example: [GPT-6.1 Sol pricing](https://aiapicostcalc.com/models/gpt-6-1-sol-pricing) and [Gemini 3.1 Pro pricing](https://aiapicostcalc.com/models/gemini-3-1-pro-preview-pricing).", "url": "https://wpnews.pro/news/the-context-cliff-200k-272k-and-the-line-that-doubles-your-prompt-bill", "canonical_source": "https://dev.to/succday/the-context-cliff-200k-272k-and-the-line-that-doubles-your-prompt-bill-3aco", "published_at": "2026-11-21 01:27:49+00:00", "updated_at": "2026-11-21 01:42:12.135535+00:00", "lang": "en", "topics": ["large-language-models", "ai-infrastructure"], "entities": ["OpenAI", "Google", "xAI", "Anthropic", "GPT-6.1 Sol", "Gemini 3.1 Pro", "Grok 4.3", "Claude Haiku 5.5"], "also_reported_by": [], "takeaway": "A 250K-token request costs $0.60 per call on GPT-6.1 Sol versus $1.18 on Gemini 3.1 Pro, because Google's 200K threshold re-prices the whole request at double the input rate while OpenAI's line sits at 272K.", "recommendation": {"advice": "Check each model's long-context threshold and whether the whole request is re-priced before budgeting; a 250K-token workload costs $600/month on GPT-6.1 Sol versus $1,180 on Gemini 3.1 Pro.", "playbook": "llm-cost", "playbook_url": "https://wpnews.pro/playbooks/llm-cost"}, "alternates": {"html": "https://wpnews.pro/news/the-context-cliff-200k-272k-and-the-line-that-doubles-your-prompt-bill", "markdown": "https://wpnews.pro/news/the-context-cliff-200k-272k-and-the-line-that-doubles-your-prompt-bill.md", "text": "https://wpnews.pro/news/the-context-cliff-200k-272k-and-the-line-that-doubles-your-prompt-bill.txt", "jsonld": "https://wpnews.pro/news/the-context-cliff-200k-272k-and-the-line-that-doubles-your-prompt-bill.jsonld"}}