First screen, captured 4 days ago. Open the video.

Claude Fable 5.2 scores 65 on Artificial Analysis, highest so far

What HN thinks

Everyone agrees the new model is brilliant, expensive, and possibly a wall, depending on which dropdown setting you picked.

Cost per task exploded
Max-effort Fable 5.2 burns far more tokens and doubles cost per task, which matters a lot for API and enterprise users who can't use subscriptions.
Lower effort settings are the story
At high, medium or low effort Fable 5.2 beats Opus 5.5 and Fable 5.1 while costing the same or less per task, and AA defaulting to max is misleading.
Diminishing returns or a wall
Some see inference-heavy scaling as a sign of hitting limits, while others argue distillation and falling cost-per-capability show progress continues.

Consensus: Most agree the max setting is rarely worth it and that cheaper models suffice for many tasks, but there's no agreement on whether frontier pricing is justified.

Recommended in the comments

GPT-6.1 Sol (product): Commenters cite it as much more economical per task than Opus with comparable results and less moralizing.
DeepSeek V4.1 Flash (product): Recommended as capable at a tiny fraction of Anthropic's price.
Opik ↗ (tool): Mentioned by a former developer in the context of measuring agent performance at scale.
OpenThoughts-114k ↗ (other): Pointed to as research on distilling long reasoning into shorter traces.

Summarised by Claude from the comment thread 3 days ago.

Points: 347
Comments: 247
Best rank: #2
Posted: 2026-11-09 by theanonymousone
Category: AI & ML

claude-fable-5-2 anthropic artificial-analysis llm-benchmarks ai-model-pricing

All captures:

Full page

Captured 4 days ago. Click for the whole screenshot.

Looks similar

Screenshot of huggingface.co: Kimi K3.1 Pro model weights are out

Kimi K3.1 Pro model weights are out (huggingface.co)

607 points by abrarthought 9 days ago | 174 comments

Hugging Face model card for Moonshot AI's Kimi K3.1 Pro, an open-weight 2.8T-parameter multimodal MoE model with 1M-token context, including architecture and benchmark results.

kimi moonshot-ai open-weights llm

Screenshot of ycombinator.com: Y/N (YC S23) is hiring a growth lead to distribute superhuman A.I. storytelling

Y/N (YC S23) is hiring a growth lead to distribute superhuman A.I. storytelling (ycombinator.com)

0 points by gabrielbirnbaum 4 days ago | 0 comments

Y Combinator job listing for a Founding Growth Marketer at Y/n (YC S23), an AI interactive storytelling app, in San Francisco paying $130K–$300K plus equity.

y-combinator hiring growth-marketing interactive-fiction

Screenshot of ycombinator.com: Mbodi AI (YC P25) Is Hiring Research and Robotic Engineers

Mbodi AI (YC P25) Is Hiring Research and Robotic Engineers (ycombinator.com)

0 points by chitianhao 7 days ago | 0 comments

Y Combinator jobs page for Mbodi AI (YC P25), an embodied AI robotics startup hiring a Founding Research Engineer and Senior Robotics Engineer in New York.

robotics embodied-ai y-combinator hiring

Screenshot of ycombinator.com: Bild AI (YC W25) Is Hiring Founding Product Engineers

Bild AI (YC W25) Is Hiring Founding Product Engineers (ycombinator.com)

0 points by rooppal 8 days ago | 0 comments

Y Combinator jobs page for Bild AI (W25), an AI estimating and detailing startup for construction doors and hardware, listing founding engineer and intern roles.

y-combinator hiring bild-ai construction-tech

Screenshot of github.com: Show HN: Flide: Make Jev-like decision models up to 9.5x faster

Show HN: Flide: Make Jev-like decision models up to 9.5x faster (github.com)

103 points by unikcc 4 days ago | 25 comments

GitHub repository for Flide, an open-source inference engine for Jev-style decision models using exact state forking, claiming up to 9.5x speedup over vLLM and SGLang.

llm-inference inference-engine decision-models jev

Screenshot of lists.trustedfirmware.org: Future Maintenance of Mbed TLS

Future Maintenance of Mbed TLS (lists.trustedfirmware.org)

61 points by jonesjohnson 10 days ago | 17 comments

Mailing list announcement that current Mbed TLS maintainers will stop maintaining the TLS and X.509 libraries from April 2027 unless new maintainers step in, with user replies.

mbed-tls tls x509 tf-psa-crypto