--- schema: 1 type: video title: >- 20 days of compute vs 7 hours: rethinking what state-of-the-art means — Bertrand Charpentier, Pruna description: >- Ranking image generation models the way Design Arena does it — 26,000 battles, 62 seconds per generation — takes 20 days of compute, costs $5,000, and consumes roughly 400 marathons worth of energy. Bertrand Charpentier, cofounder and chief scientist at Pruna AI, uses that number to make a point: the same evaluation on a fast compressed model takes 7 hours and $265. Efficiency is a dimension of state of the art, not a footnote. The rest of the talk dismantles the idea that any single model holds the title. Leaderboard rankings disagree with each other — the same model goes from rank 10 on one to rank 5 on another. Most models lose 40% of their head-to-head battles, which means the top-ranked model is the wrong choice for nearly half of real use cases. His answer is the Pareto front: plot quality against latency or cost, find the frontier, and expect three or four models clustered tightly in quality score but varying up to 20x in efficiency. Evaluating this way tends to surface small specialized models rather than large foundation models. Speaker info: - https://www.linkedin.com/in/bertrand-charpentier-76995ab6/ - https://github.com/sharpenb --- Watch along with the corrected transcript, summary, timestamps and resources on this talk's official AIE page: https://ai.engineer/talks/hqHC6Z_lXyo-20-days-of-compute-vs-7-hours-rethinking-what-state-of-the-art-means-bertrand-charpentier-pruna Join our upcoming AI Engineer conferences (see https://ai.engineer landing page): - AIEi Shanghai, November 5–6 — the first for meeting top Chinese labs and engineers - AIE CODE in SF, November 10–12 — the top AI Coding conference returns - AIEi Sydney, December 7–8 — the first AIEi run alongside NeurIPS 2026 in Sydney --- canonical: >- https://liquidgold.cc/channel/aidotengineer/explore/video/20-days-of-compute-vs-7-hours-rethinking-what-state-749de1a184f5 markdown: >- https://liquidgold.cc/channel/aidotengineer/video/20-days-of-compute-vs-7-hours-rethinking-what-state-749de1a184f5.md llms: null parent: 'https://liquidgold.cc/channel/aidotengineer.md' channel: aidotengineer updated: '2026-10-24T10:06:10.456Z' language: null children: [] videoId: hqHC6Z_lXyo durationSeconds: 1176 publishedAt: '2026-06-01T17:00:06.000Z' searchChunks: - ordinal: 0 text: null startSeconds: 25 endSeconds: 144 watch: 'https://liquidgold.cc/watch/hqHC6Z_lXyo?t=25' heldReason: rights - ordinal: 1 text: null startSeconds: 155 endSeconds: 287 watch: 'https://liquidgold.cc/watch/hqHC6Z_lXyo?t=155' heldReason: rights - ordinal: 2 text: null startSeconds: 287 endSeconds: 298 watch: 'https://liquidgold.cc/watch/hqHC6Z_lXyo?t=287' heldReason: rights - ordinal: 3 text: null startSeconds: 298 endSeconds: 378 watch: 'https://liquidgold.cc/watch/hqHC6Z_lXyo?t=298' heldReason: rights - ordinal: 4 text: null startSeconds: 390 endSeconds: 482 watch: 'https://liquidgold.cc/watch/hqHC6Z_lXyo?t=390' heldReason: rights - ordinal: 5 text: null startSeconds: 494 endSeconds: 605 watch: 'https://liquidgold.cc/watch/hqHC6Z_lXyo?t=494' heldReason: rights - ordinal: 6 text: null startSeconds: 617 endSeconds: 716 watch: 'https://liquidgold.cc/watch/hqHC6Z_lXyo?t=617' heldReason: rights - ordinal: 7 text: null startSeconds: 716 endSeconds: 728 watch: 'https://liquidgold.cc/watch/hqHC6Z_lXyo?t=716' heldReason: rights - ordinal: 8 text: null startSeconds: 728 endSeconds: 808 watch: 'https://liquidgold.cc/watch/hqHC6Z_lXyo?t=728' heldReason: rights - ordinal: 9 text: null startSeconds: 818 endSeconds: 940 watch: 'https://liquidgold.cc/watch/hqHC6Z_lXyo?t=818' heldReason: rights - ordinal: 10 text: null startSeconds: 940 endSeconds: 951 watch: 'https://liquidgold.cc/watch/hqHC6Z_lXyo?t=940' heldReason: rights - ordinal: 11 text: null startSeconds: 951 endSeconds: 1072 watch: 'https://liquidgold.cc/watch/hqHC6Z_lXyo?t=951' heldReason: rights - ordinal: 12 text: null startSeconds: 1082 endSeconds: 1176 watch: 'https://liquidgold.cc/watch/hqHC6Z_lXyo?t=1082' heldReason: rights - ordinal: 13 text: null startSeconds: 1082 endSeconds: 1176 watch: 'https://liquidgold.cc/watch/hqHC6Z_lXyo?t=1082' heldReason: rights --- # 20 days of compute vs 7 hours: rethinking what state-of-the-art means — Bertrand Charpentier, Pruna Ranking image generation models the way Design Arena does it — 26,000 battles, 62 seconds per generation — takes 20 days of compute, costs $5,000, and consumes roughly 400 marathons worth of energy. Bertrand Charpentier, cofounder and chief scientist at Pruna AI, uses that number to make a point: the same evaluation on a fast compressed model takes 7 hours and $265. Efficiency is a dimension of state of the art, not a footnote. The rest of the talk dismantles the idea that any single model holds the title. Leaderboard rankings disagree with each other — the same model goes from rank 10 on one to rank 5 on another. Most models lose 40% of their head-to-head battles, which means the top-ranked model is the wrong choice for nearly half of real use cases. His answer is the Pareto front: plot quality against latency or cost, find the frontier, and expect three or four models clustered tightly in quality score but varying up to 20x in efficiency. Evaluating this way tends to surface small specialized models rather than large foundation models. Speaker info: - https://www.linkedin.com/in/bertrand-charpentier-76995ab6/ - https://github.com/sharpenb --- Watch along with the corrected transcript, summary, timestamps and resources on this talk's official AIE page: https://ai.engineer/talks/hqHC6Z_lXyo-20-days-of-compute-vs-7-hours-rethinking-what-state-of-the-art-means-bertrand-charpentier-pruna Join our upcoming AI Engineer conferences (see https://ai.engineer landing page): - AIEi Shanghai, November 5–6 — the first for meeting top Chinese labs and engineers - AIE CODE in SF, November 10–12 — the top AI Coding conference returns - AIEi Sydney, December 7–8 — the first AIEi run alongside NeurIPS 2026 in Sydney --- Canonical: https://liquidgold.cc/channel/aidotengineer/explore/video/20-days-of-compute-vs-7-hours-rethinking-what-state-749de1a184f5 · Channel: [AI Engineer](https://liquidgold.cc/channel/aidotengineer.md) > Search chunk text is held until public-source provenance is verified. ## Document chunks > Missing: no stored document chunks are available.