Twitxy Saved

Physion Labs Official

Verified account This account is verified. Learn more

@Physion_Labs

Physion Labs is building the trust layer for video generation and world models.
Bellevue, WA physionlabs.ai Joined July 2023
6 Following 493 Followers 47 Posts

Latest posts 5

Pinned post

Physion Labs Official Verified account @Physion_Labs
The AI industry has a massive compute efficiency problem. At Physion Labs, we think it's time to challenge it.

Today, we're sharing the technology behind NexTron, our world-intelligence training engine.

Our benchmarks on matched hardware:
⚡ Up to 20.9× SFT throughput vs. DeepSpeed (H100)
⚡ Up to 6.6× SFT throughput vs. Megatron (H200)
⚡ Up to 7.2× SFT throughput vs. Megatron (B200)
⚡ Up to 3.1× end-to-end RL throughput vs. Miles–Megatron
🔥 180B model fine-tuning on just 8 H100 GPUs.

Same hardware. Dramatically higher throughput. Comparable model quality.

We believe the next frontier isn't just building bigger GPU clusters.

It's unlocking more intelligence from every GPU, every watt, and every dollar spent on training.

#PhysionLabs #NexTron #WorldIntelligence #AIInfrastructure
14 replies 47 reposts 99 likes 101K views View on X
Physion Labs Official Verified account @Physion_Labs
Physion-Arc 1.0 Ads & Marketing Leaderboard Update: physionlabs.ai/leaderboards/a…! Evaluated across 100 real-world advertising briefs and 24 metrics.

Excited to add @frameo_ai Manto 2.0 and @higgsfield_ai Supercomputer to the leaderboard.

🏆 Latest rankings:
🥇 @frameo_ai Manto 2.0 — 71.8
🥈 @runwayml Agent 2.0 — 68.6
🥉 @openart_ai — 66.8
4. @higgsfield Supercomputer — 66.6
5. @Creatify_AI Ads Agent — 62.2
6. @LumaLabsAI Creative Agents — 59.2
7. @HeyGen Video Agent — 51.0

Great to see continued progress across AI advertising agents!
Physion Labs Official Verified account @Physion_Labs
Who's building the best AI avatar video agent? 🏆

We just released the Physion-Arc 1.0 Avatar Track, independently benchmarking 7 leading AI video agents on 100 real-world creative briefs.

The results are in:
🥇 @higgsfield_ai Supercomputer — 75.3
🥇 @HeyGen Video Agent — 73.1
🥉@runwayml Agent 2.0 — 72.1
4️⃣ @Creatify_AI AI Creative Agent — 71.4
5️⃣ @fal Agent — 69.4
6️⃣ @LumaLabsAI Creative Agents — 64.0
7️⃣ @Flovaai Agent — 61.8

@higgsfield_ai and @HeyGen statistically tie for #1, with Higgsfield holding a slight 2.2-point lead that is not statistically significant.

Every agent was evaluated by human annotators across 19 metrics, covering content fidelity, production quality, and human realism — including multilingual consistency.

AI video is evolving from generating clips to autonomously producing finished content. We believe benchmarking needs to evolve with it.

Congratulations to all seven teams pushing the frontier forward! 🚀

🔗 Full leaderboard: physionlabs.ai/leaderboards/a…
6 replies 4 reposts 26 likes 4.6K views View on X
Physion Labs Official Verified account @Physion_Labs
Great to see @Physion_Labs represented at @MIT alongside @hanrui_w and @LLiebenwein for a conversation on where AI infrastructure is headed and what remains hard to build.

Thank you to @e14fund, @c10n, and Habib Haddad for bringing together founders, researchers, and students for the discussion. We’re excited to keep pushing the frontier of AI infrastructure and world understanding — and to meet the next generation of researchers and builders along the way 🚀🚀🚀
0 replies 2 reposts 10 likes 8.3K views View on X
Physion Labs Official Verified account @Physion_Labs
🚀 𝐏𝐡𝐲𝐬𝐢𝐨𝐧-𝐀𝐫𝐜 1.0 𝐀𝐝𝐬 & 𝐌𝐚𝐫𝐤𝐞𝐭𝐢𝐧𝐠 𝐓𝐫𝐚𝐜𝐤 𝐢𝐬 𝐥𝐢𝐯𝐞.

For this track, we wanted to test something very practical: can today’s video agents actually take a real advertising brief and turn it into a strong finished ad?

We worked with advertising and creative professionals to build 100 real-world briefs across text-only, single-reference, and multi-reference tasks, then evaluated each agent across 24 metrics covering brief fidelity, production quality, and ad effectiveness.

The first leaderboard includes @runwayml, @openart_ai, @Creatify_AI, @LumaLabsAI, and @HeyGen.

One result we found especially interesting: Runway and OpenArt AI are very close on overall quality — 71.1 vs. 69.5, with no statistically significant difference in this round.

💰 At the same time, the agents show quite different cost profiles. OpenArt AI is about $6/video, compared with $37/video for Runway, which makes the quality–cost tradeoff another interesting dimension to look at alongside overall performance.

Full leaderboard here: physionlabs.ai/leaderboards/a…