A universal cure or vaccine effectively eliminating cancer mortality across most major types is approved by regulatory agencies before 2040.
Oncologists debate whether cancer can ever have a single 'cure' given its genetic diversity. Resolution requires FDA or EMA approval of a broadly curative therapy for most cancers.
Across 6 frontier models answering independently, the panel has all but ruled it out: the median is 5%.
They broadly agree, 5 points between GPT at 7% and Claude at 2%.
Nothing is settled until 12 Sept 2076, when the outcome is checked against sources and every number above gets a permanent score.
What each one said
-
GPTOpenAI
Cancer’s biological diversity makes one regulator-approved intervention spanning most lethal types a very high bar.
- FDA or EMA would need pivotal trials showing durable mortality elimination across disparate solid and blood cancers, not subtype gains.
- No prior therapy or vaccine has produced pan-cancer cures; checkpoint inhibitors and targeted drugs benefit selected tumors and patients.
- A validated shared cancer vulnerability with broad, durable human results before 2040 would sharply raise the odds.
7% -
GrokxAI
Turns on inventing, trialing and approving a pan-cancer therapy that cuts mortality across most types in <14 years.
- FDA/EMA require multi-year Phase 3 data proving mortality reduction in lung, breast, colorectal, prostate and similar major cancers.
- No pan-cancer cure has emerged in 50 years of oncology; genetic diversity keeps this well below that already-low base rate.
- A 2027-28 universal neoantigen or mRNA vaccine with early mortality signals would shift odds, but none is in late trials now.
7% -
DeepSeekDeepSeek
Cancer is hundreds of genetically distinct diseases; a single broadly curative approval by 2040 is a moonshot, not a pipeline event.
- FDA and EMA approve products for specific indications, so a 'most major types' label would require either a pan-tumor mechanism or dozens of separate approvals.
- The closest precedents — checkpoint inhibitors, CAR-T, mRNA vaccines — each cover narrow subsets and extend survival rather than eliminate mortality.
- Base rate for a genuine pan-cancer cure in 14 years is near zero; mRNA neoantigen vaccines and early-detection blood tests are the only plausible bridges, and neither is…
6% -
GeminiGoogle
Cancer comprises hundreds of distinct molecular diseases, making a single universal curative mechanism biologically implausible by 2040.
- FDA approval requires Phase III trials across diverse tumors, an operational process taking 10-15 years once a broad target is found.
- Pan-cancer mechanisms like p53 targeting or telomerase inhibition have repeatedly failed to yield universal, non-toxic efficacy.
- Discovery of a truly ubiquitous neoantigen or metabolic vulnerability that spares healthy tissue would dramatically raise odds.
3% -
ClaudeAnthropic
Cancer's biological diversity makes a single universal cure scientifically implausible, even with 50 years of runway.
- Cancer isn't one disease but thousands of distinct molecular pathologies, so no single mechanism could cure 'most major types'.
- Current trends favor precision oncology (tumor-specific therapies) over universal cures, and no candidate approach exists today.
- Even dramatic mortality reductions via early detection or targeted therapy wouldn't satisfy a literal 'universal cure or vaccine' bar for regulatory approval.
2% -
QwenAlibaba
Cancer is hundreds of diseases; no single therapy has ever cleared even one major type universally.
- FDA/EMA have never approved a therapy curative across multiple solid tumor types; CAR-T and checkpoint inhibitors remain type-specific.
- Even the most optimistic pipeline (mRNA neoantigen vaccines, pan-cancer TCR therapies) is in Phase I/II with no Phase III data before 2030.
- A development that would shift this: a single agent showing >90% complete response across 5+ tumor types in a pivotal trial, which has no precedent.
2%
6 of 6 models answered · 5 points between the highest and lowest. None was shown the market price.
Question sourced from a news sweep on 25 Sept 2026. Forecast by google/gemini-3.8-flash, anthropic/claude-sonnet-5, openai/gpt-5.6-terra, x-ai/grok-4.6, deepseek/deepseek-v4.1-flash, qwen/qwen3.8-max-0902 via OpenRouter.