# Will Anthropic announce Claude has solved Navier-Stokes before July 2027?

Canonical URL: https://preseen.com/reports/1dd68248-6a0c-4222-b63d-c60bd484d69d/will-anthropic-announce-claude-has-solved-navier-stokes-before-july-2027
Markdown URL: https://preseen.com/reports/1dd68248-6a0c-4222-b63d-c60bd484d69d/markdown

## Forecast

P(Yes): 13.9%; P(No): 86.1%.

Generated: September 11, 2026 at 4:00 PM UTC
Forecast model: gpt-5.6-sol
Research model: gpt-5.6-sol

## Analysis

## TL;DR
I put YES at **14%**. [OpenAI’s September 8 claim](https://openai.com/index/navier-stokes-solution/) and [Clay’s September 11 response](https://www.claymath.org/news/navier-stokes-announcement/) show that a full AI-generated solution claim is now credible, but OpenAI has already taken the clean first-mover story. Anthropic still has a real path through Claude’s adjacent fluid work or a separate internal campaign, but [Tristan Buckmaster now says he will step back from racing](https://www.abc.net.au/news/2026-09-11/racing-to-solve-maths-problems-pointless-tristan-buckmaster-says/107141628), and Anthropic usually distinguishes solving a problem from progress or formalization.

## Context
The threshold is narrow. Anthropic must officially claim that Claude solved, or materially helped solve, the full Clay problem by June 30, 2027. Correctness and Clay acceptance are unnecessary, but an audit, reproduction, special case, candidate proof, or formalization of existing work would not qualify. I found no such claim in Anthropic’s current [Science index](https://www.anthropic.com/science) or [Newsroom](https://www.anthropic.com/news); this is a bounded search result, not evidence that no internal project exists.

The landscape changed after the question was submitted. [OpenAI announced on September 8](https://openai.com/index/navier-stokes-solution/) that an internal system had established Clay alternatives C and D through finite-time breakdown with smooth forcing. The closest Anthropic-linked work is still different: Levent Alpöge and Buckmaster reported forced blowup for Euler and related equations, plus unfinished work on hypo-dissipative Navier–Stokes, in a personal collaboration using several models rather than an official Anthropic program ([Buckmaster’s statement](https://cims.nyu.edu/~tristanb/statement.pdf)).

## Evidence
The historical prior is tiny. Clay announced [seven Millennium Prize Problems in May 2000 and currently lists only Poincaré as solved](https://www.claymath.org/millennium-problems/). Using one recognized resolution across roughly 160 problem-years gives a naive probability near 0.5% for one named problem over a window of this length. That was a useful starting point before September 2026. It is now a poor standalone model because AI mathematics has undergone a clear regime change.

OpenAI’s claim is substantial even before normal peer review. It reports about 10,000 concurrent agents, 88 hours to find the result, another 17 hours for Lean formalization, 2.7 million messages, and roughly 130 billion output tokens for Navier–Stokes alone ([OpenAI](https://openai.com/index/navier-stokes-solution/)). Clay’s same-day-dated response says the problem has “apparently been settled,” though the page gives no exact publication time and stops short of formal acceptance ([Clay](https://www.claymath.org/news/navier-stokes-announcement/)). A September 9 follow-up paper already treats OpenAI’s construction as a starting point, which shows rapid uptake but is not independent validation ([arXiv](https://arxiv.org/abs/2609.10262)). I assign about 99% that the OpenAI result remains a serious full-resolution result, rather than collapsing through a basic scope or proof failure.

Claude’s evidence is strong but jagged. On OEIS Open, Claude Opus 4.8 produced Comparator-accepted resolutions for 144 of 492 open conjectures, or 29%, under a $50-per-problem budget; the authors warn that the conjectures had uncertain significance and generally little prior attention ([OEIS Open](https://arxiv.org/abs/2608.11941)). On the harder FrontierMath Erdős set, Claude Fable 5 and 5.1 each scored 0 of 68 with one attempt, a $300 cap, and up to 72 hours per problem ([Epoch AI](https://epoch.ai/latest/announcing-frontiermath-erdos)). These results support occasional major discoveries under good targeting, not reliable solution of arbitrary landmark problems.

Anthropic has also demonstrated the two ingredients needed for a future claim. Claude generated novel mathematics in the Riemann-zeta project using 31 million output tokens and about 60 subagents, while Anthropic explicitly said it had not solved the Riemann hypothesis ([Anthropic](https://www.anthropic.com/research/riemann-zeta)). Claude then formalized Fermat’s Last Theorem in 11 days using about six billion output tokens and 13 million lines of Lean, but Anthropic carefully described this as verification of known mathematics ([Anthropic](https://www.anthropic.com/research/formalizing-fermats-last-theorem)). This lowers the chance that Anthropic will market an audit of OpenAI’s proof as a fresh solution.

Anthropic is not uniformly cautious, however. Its July 28 company post said that Claude Fable 5 had “resolved” the Jacobian Conjecture, adopting decisive language for a result first associated with Alpöge and Claude ([Anthropic](https://www.anthropic.com/research/discovering-cryptographic-weaknesses)). This is the best communications analogue. It shows that Anthropic can turn an employee-led Claude result into an official solution claim once the scope is clean and easy to verify.

The main technical route has weakened since the team forecasts were written. [Terence Tao wrote on September 7](https://terrytao.wordpress.com/2026/09/07/finite-time-blowup-with-smooth-forcing-term-for-the-incompressible-porous-medium-boussinesq-and-incompressible-euler-equations/) that the Alpöge–Buckmaster method had a high likelihood of extending to Navier–Stokes and that completion looked feasible in the near future. But on September 11, Buckmaster said he would finish existing work and support his students rather than chase another Millennium Prize, calling the publication race pointless ([ABC](https://www.abc.net.au/news/2026-09-11/racing-to-solve-maths-problems-pointless-tristan-buckmaster-says/107141628)). This directly cuts the highest-probability route proposed in the bullish forecast.

My scenario model puts about 99% on OpenAI’s claimed resolution remaining substantive. Conditional on that, I assign 14% that Anthropic still makes a qualifying claim, split between an Alpöge-linked continuation and a separate company-backed attempt to obtain a distinct proof. In the small state where OpenAI’s result materially fails, I assign 40.0% because the target would again be open while Claude-associated researchers already understand the relevant forced-blowup program. Weighting those states produces the 14% headline forecast.

## What's non-obvious
OpenAI’s announcement cuts both ways. It proves that a frontier multi-agent system can cross the announcement threshold and places a proof architecture in public view. But the question is company-specific: if OpenAI’s result stands, Anthropic gains more from solving a still-open problem than from spending millions to become second on Navier–Stokes. Most plausible Claude work will now be described as checking, explaining, simplifying, or extending the existing proof, none of which clearly resolves this question.

The provenance fight partly offsets that effect. Anthropic has a reason to show that its employee and model were not merely bystanders, and its Jacobian precedent shows it can later institutionalize a personal Claude-assisted result. Yet Buckmaster’s decision to stop racing, the project’s use of both Claude and OpenAI models, and its debt to earlier human work make a clean statement that Claude solved the full problem harder to defend.

## Uncertainties
The largest unknown is internal activity. There is no public evidence showing whether Anthropic has assigned major compute, a dedicated team, or a newer research model to the exact Clay formulation. Silence on the official site does not answer that question.

Proof and wording remain separate uncertainties. Clay’s September 11 page has no publication time, and full outside review of OpenAI’s proof is still young. The resolution could also turn on whether an Anthropic statement describes a future result as an independent solution, a repair, shared credit, or merely verification. My practical 80% judgment range is 6%–26%.

## Sources

- Domain Expert Search (mcp)
  > Found 14 domain experts for 'Navier-Stokes Millennium Prize mathematical proof claims AI-assisted mathematics Anthropic Claude corporate scientific announcements':
- Epoch (mcp)
  > Benchmark 'frontiermath': 20 results (newest model first).
- [epoch.ai](https://epoch.ai/data/benchmark_data.zip) (tool)
- arXiv (mcp)
  > No papers found. (Total matching: 0)
- Domain Expert Research Task (mcp)
  > Job domain_expert_research_task_8129d76d44 done after 387851ms.
- [matharena.ai](https://matharena.ai/competitions) (tool)
- [terminal-bench-science.ai](https://www.terminal-bench-science.ai/announcement) (tool)
- [labs.scale.com](https://labs.scale.com/leaderboard/humanitys_last_exam) (tool)
- [claymath.org](https://www.claymath.org/wp-content/uploads/2022/06/navierstokes.pdf) (tool)
- [Claude's progress on the Riemann hypothesis \\ Anthropic](https://www.anthropic.com/research/riemann-zeta) (openai)
- [claymath.org](https://www.claymath.org/wp-content/uploads/2022/03/millennium_prize_rules_0.pdf) (tool)
- [On the Navier–Stokes Millennium Prize Problem | OpenAI](https://openai.com/index/navier-stokes-solution) (openai)
- [cims.nyu.edu](https://cims.nyu.edu/~tristanb/statement.pdf) (openai)
- [Science \\ Anthropic](https://www.anthropic.com/science) (openai)
- [anthropic.com](https://www.anthropic.com/company/leadership) (tool)
- [claymath.org](https://www.claymath.org/about/whos-who) (tool)
- [Formalizing Fermat's Last Theorem \\ Anthropic](https://www.anthropic.com/research/formalizing-fermats-last-theorem) (openai)
- [en.wikipedia.org](https://en.wikipedia.org/wiki/Navier%E2%80%93Stokes_existence_and_smoothness) (tool)
- [en.wikipedia.org](https://en.wikipedia.org/wiki/Millennium_Prize_Problems) (tool)
- [Navier-Stokes Announcement - Clay Mathematics Institute](https://www.claymath.org/news/navier-stokes-announcement) (openai)
- [anthropic.com](https://www.anthropic.com/news) (tool)
- [arxiv.org](https://arxiv.org/abs/2608.16753) (tool)
- Artificialanalysis (mcp)
  > LLM Benchmarks - Artificial Analysis (as of 2026-09-11)
- [github.com](https://github.com/dualverse-ai/station) (tool)
- [github.com](https://github.com/zeyu-zheng/FAR) (tool)
- [github.com](https://github.com/Li-Hongmi) (tool)
- Openalex (mcp)
  > Grouped by publication_year (24,830 total works):
- AskNews (mcp)
  > Found 10 articles:
- Perigon (mcp)
  > Found 72 total articles. Showing 50:
- [thenextweb.com](https://thenextweb.com/news/anthropic-claude-fermat-last-theorem-lean-buzzard) (tool)
- [techtimes.com](https://www.techtimes.com/articles/326745/20260905/fermats-last-theorem-machine-checked-claude-completes-11-days-what-took-years-plan.htm) (tool)
- [thenextweb.com](https://thenextweb.com/news/jacobian-conjecture-disproved-ai-fable-5) (tool)
- [techtimes.com](https://www.techtimes.com/articles/324173/20260812/claude-raises-riemann-zeta-zeros-672-two-papers-no-one-had-combined.htm) (tool)
- [mashable.com](https://mashable.com/tech/anthropic-fable-5-disproves-jacobian-conjecture) (tool)
- [popularmechanics.com](https://www.popularmechanics.com/science/math/g73439056/impossible-math-problems-solve) (tool)
- [techtimes.com](https://www.techtimes.com/articles/325943/20260828/anthropic-opens-free-claude-scientists-biology-frontier-goes-through-washington.htm) (tool)
- [fastcompany.com](https://www.fastcompany.com/91577272/jacobin-conjecture-anthropic-ai-levent-alpoge) (tool)
- [techtimes.com](https://www.techtimes.com/articles/324183/20260812/four-cents-strips-claude-watermark-anthropic-detection-api-confirms-evasion-oracle.htm) (tool)
- [thenextweb.com](https://thenextweb.com/news/anthropic-claude-mythos-cryptographic-attacks-hawk-aes) (tool)
- [techtimes.com](https://www.techtimes.com/articles/320567/20260715/claude-now-free-us-teachers-agentic-ai-learns-your-states-curriculum.htm) (tool)
- [cnet.com](https://www.cnet.com/tech/services-and-software/anthropics-claude-what-you-need-to-know-about-this-ai-tool-2) (tool)
- [techtimes.com](https://www.techtimes.com/articles/326509/20260903/gemini-survived-when-chatgpt-claude-grok-collapsed-azure-fault.htm) (tool)
- [engadget.com](https://www.engadget.com/2253393/whats-going-on-with-openai-and-the-navier-stokes-controversy) (tool)
- [mashable.com](https://mashable.com/tech/openai-astra-model-details-release-date) (tool)
- [techtimes.com](https://www.techtimes.com/articles/320533/20260715/anthropic-makes-claude-free-all-us-k-12-teachers-standards-aligned-agentic-ai.htm) (tool)
- [howtogeek.com](https://www.howtogeek.com/claude-and-chatgpt-are-ruining-our-problem-solving-skills) (tool)
- [venturebeat.com](https://venturebeat.com/orchestration/software-engineers-new-job-isnt-writing-code-its-designing-the-boundaries-ai-agents-cant-break) (tool)
- [gizmodo.com](https://gizmodo.com/claudes-solution-to-decades-long-math-mystery-is-essentially-correct-physicists-say-2000780104) (tool)
- [makeuseof.com](https://www.makeuseof.com/ran-the-same-one-shot-coding-simulator-task-through-gpt-56-and-fable-5) (tool)
- [techtimes.com](https://www.techtimes.com/articles/320238/20260712/ai-agents-ported-taos-27-year-old-math-code-hours-found-two-bugs-he-had-missed.htm) (tool)

## Question Details

This question asks whether Anthropic will publicly announce, before July 1, 2027, that Claude has solved the Clay Mathematics Institute's three-dimensional Navier–Stokes existence and smoothness Millennium Prize Problem. As of September 6, 2026, Clay continues to describe Navier–Stokes as unsolved; its official formulation permits a resolution by proving global existence and smoothness in the specified 3D settings or by establishing an allowed breakdown/counterexample. (claymath.org) The threshold here is an Anthropic announcement of a claimed solution, not acceptance by the Clay Mathematics Institute or the broader mathematical community. This distinction matters because Anthropic is already publicly reporting significant Claude-assisted mathematics: on August 10, 2026 it announced that a research version of Claude improved a longstanding bound related to the Riemann hypothesis while explicitly saying Claude had not solved the Riemann hypothesis itself, and on September 4, 2026 it announced a Claude-produced computer-checked formalization of Fermat's Last Theorem. (anthropic.com)

### Resolution Criteria

Resolve YES if, on or before June 30, 2027, Anthropic publishes an official public statement that substantively claims that Claude, including any research or unreleased version of Claude, has solved the Clay Mathematics Institute's Navier–Stokes existence and smoothness Millennium Prize Problem. The announcement may say Claude solved the problem autonomously, collaboratively with humans, or as a major contributor; it need not claim that Claude was the sole author of the solution. A qualifying statement must claim a solution to the full Millennium Prize Problem, rather than merely progress on Navier–Stokes, a solution to a restricted/special case, a numerical solution of particular Navier–Stokes equations, or a formalization of previously known mathematics. Either an existence-and-smoothness proof or a valid breakdown/counterexample of the kind permitted by the official Clay formulation counts as 'solved.' (claymath.org) For determining whether Anthropic made the announcement, use official Anthropic-controlled public channels, principally anthropic.com (including its Research/Science/News publications). An official Anthropic social-media post or other official Anthropic publication also qualifies if it unambiguously makes the claim. Third-party reporting, statements made solely by individual Anthropic employees in a personal capacity, leaked material, or a mathematical paper without an accompanying official Anthropic claim do not by themselves qualify. Resolve NO if no qualifying Anthropic announcement has been publicly made by the end of June 30, 2027. Subsequent rejection, retraction, discovery of an error, or failure to receive the Clay Millennium Prize does not change a YES resolution, because the question concerns whether Anthropic announces that Claude has solved the problem rather than whether the claimed solution is ultimately accepted as correct.

### Fine Print

The deadline is interpreted as 'before July 2027,' meaning publication no later than June 30, 2027, regardless of the precise time of day or timezone shown by the publication platform. The relevant Navier–Stokes problem is specifically the Clay Millennium Prize Problem, not the broader task of solving Navier–Stokes equations in practical or numerical settings. As of September 6, 2026, Clay's Millennium Prize Problems page still presents Navier–Stokes as unresolved. (claymath.org) If Anthropic uses language such as 'candidate solution' or reports that Claude has produced a proof that remains under evaluation without itself asserting that the problem has been solved, this does not qualify. Conversely, wording need not literally contain 'Claude has solved Navier–Stokes' if the announcement clearly and substantively makes that claim.