Users compare Fable 5.1 and GPT-6 Astra across coding and asset tasks
Side-by-side tests on Reddit pit Fable 5.1 against GPT-6 Astra (both on xhigh): one 2D sprite prompt yielded 16 key poses from Astra versus 992 frames across four palettes plus a Python generator from Fable. A separate ML workflow test found Astra codes more agentically while Fable writes and follows directions better.
How this story unfolded
3 days · 0 reports · 7 community posts · from Sep 5
- Sep 5
- Sep 6
- Sep 8
More stories today
OpenAI launches ChatGPT for Financial Services
The tailored ChatGPT Work experience combines built-in financial data with GPT-6 Astra's reasoning for research, financial modeling, and client-ready materials. CNBC frames it as targeting the labor-intensive pitchbook and modeling work traditionally handled by junior bankers.
OpenAI Blog·47 minutes ago
Researchers debate what remains for scientists as LLMs take on research tasks
Anna Rogers·56 minutes agoTogether AI launches preemptible GPU compute at half price
Together GPU Clusters now offers preemptible nodes at a flat 50% of the on-demand rate, billed sub-hourly with usage metered every one to two minutes. Reclaimed nodes get a five-minute drain window (SIGTERM at T+0, removal at T+5:00), and clusters auto-refill toward the preemptible target.
Together AI Blog·57 minutes ago
PixVerse V6 enters Text-to-Video Arena at #20
PixVerse V6 debuted at #20 on Artificial Analysis's Text-to-Video Arena with a score of 1209.9. The leaderboard's top spot in the snapshot went to Gemini Omni Flash (Google) at 1324.9, ahead of Dreamina Seedance 2.0 720p (ByteDance Seed) at 1267.8.
Artificial Analysis Leaderboard·1 hour ago
PixVerse V5.6 enters Text-to-Video Arena at #19
PixVerse V5.6 debuted at #19 on Artificial Analysis's Text-to-Video Arena with a score of 1211.4. The leaderboard is topped by Gemini Omni Flash (1324.9), Dreamina Seedance 2.0 720p (1267.8) and HappyHorse-1.1 (1261.4).
Veo 3 enters Text-to-Video Arena at #17
Google's Veo 3 debuted at rank #17 with a score of 1211.9 on Artificial Analysis's blind-vote Text-to-Video Arena. Gemini Omni Flash (Google) leads the snapshot at #2 with 1324.9, ahead of Dreamina Seedance 2.0 720p (1267.8) and HappyHorse-1.1 (1261.4).
Runway Gen-4.5 debuts at #16 on Text-to-Video Arena
Runway Gen-4.5 entered Artificial Analysis's Text-to-Video Arena at #16 with a score of 1212.3. The leaderboard's top spot in the snapshot is held by Gemini Omni Flash (Google) at 1324.9, followed by Dreamina Seedance 2.0 720p (ByteDance Seed) at 1267.8.
Bach-1.0 Preview enters top 11 on Text-to-Video Arena
Video Rebirth's Bach-1.0 Preview debuted at #11 on Artificial Analysis's Text-to-Video Arena with a score of 1219.6. Gemini Omni Flash (Google) leads the snapshot at 1324.9, ahead of Dreamina Seedance 2.0 720p (1267.8) and HappyHorse-1.1 (1261.4).