Humor Arena – Which frontier model is funniest?
What if you could measure humor? Well we've trained a model on our own dataset of ~50k human ratings to detect what jokes people find funniest. We know it's part objective, part subjective component. Subjective is out of our depth for now haha The main results: Fable 5 is funniest - beating the average model 67% of the time, with GPT 4o last at 17%. Other findings: - The models never refused to try, even with dark prompts - Thinking longer has a slight benefit - Absurdness correlates negatively with joke quality Some methodology notes: - We benchmarked our model against the human majority…
In plain words
Humor Arena is a benchmark that measures which AI models generate the funniest jokes using a dataset of approximately 50,000 human ratings. It evaluates frontier models' humor capabilities by comparing their joke output against human preferences, finding that Fable 5 outperforms others significantly while GPT-4o ranks lowest. The project includes findings on how thinking time, absurdity levels, and model behavior affect joke quality. Users can rate jokes themselves on the platform to contribute to ongoing assessment.
written from the facts on this page · September 2026
From the sources
In the maker’s words, at launch
What if you could measure humor? Well we've trained a model on our own dataset of ~50k human ratings to detect what jokes people find funniest. We know it's part objective, part subjective component. Subjective is out of our depth for now haha The main results: Fable 5 is funniest - beating the average model 67% of the time, with GPT 4o last at 17%. Other findings: - The models never refused to try, even with dark prompts - Thinking longer has a slight benefit - Absurdness correlates negatively with joke quality Some methodology notes: - We benchmarked our model against the human majority and it agreed 72% of the time in a blind sample test. - We had 51 US adults rate the jokes, each blind to the models, with joke order randomized, and quality checked for attention and speed. - To rate some yourself visit https://pair.laugh.so The full benchmark here: https://laugh.so/benchmark Am taking requests if there's more research you want to see! Cheers
Does the same job
all alternatives →- EFEcho – Fable-level results at 1/3 the cost using open-weight modelsJul 2026 · ▲484
I’ve been building Echo (https://echo.tracerml.ai/), an experiment in making one AI system out of a pool of open-weight models rather than choosing a single model and using it for every task. It started with a simple experiment. I took a group of models, including GLM-5.2, Kimi K2.7 and others, and ran them on the same evaluations. Then I measured what would happen if, for each problem, you somehow knew in advance which models would be useful and how their outputs should be combined. That hypothetical system performed substantially better than any individual model in the pool.…
- OAOptimize and serve models with Fable quality at half the costJul 2026 · github.com · ▲71
Hi HN, we built world-model-optimizer, an open source tool to continually improve a specialized model for an agent. It does this by simulating production tool responses through text world modeling (similar to QwenAgentWorld, summary here https://x.com/silennai/status/2073887455884058814). We can then use this to train a router for frontier, OS, and local models (use defaults or pick which ones to optimize against). wmo ingests agent traces, builds the simulation, embeds the traces, runs different models you choose against the simulation scenarios, and then uses a KNN…
- FTFine-tune an 8B model on a 4 GB laptop GPUAug 2026 · github.com · ▲139
Is AI Dumber Today? An index of AI model experience from user's opinion23d ago · isaidumber.today · ▲17Track how AI models feel in everyday use through public community feedback, 7-day experience scores and trends. This is not a capability benchmark.
More ai this month
the category →
I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15). The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device. The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.
AI · 17d ago · simedw.com
Astute▲585Automate your B2B brand going viral, with new media creators
AI · 18d ago · company-app.joinastute.com
Hey HN, Henry from Cactus here! We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now incorporated the suggestions to release Needle 2. The whole model is a single 14MB binary that runs a full session in 28MB of RAM; 45m parameters at 2bit compression. Needle hits 500 tokens/sec decode speed on a Raspberry Pi 5, sits between 400-1,500 tokens/sec on VR devices like Meta Quest 3S and Apple Vision Pro, and ranges…
AI · 27d ago · cactuscompute.com
Launched alongside, August 2026
the whole month →- TL
Life & fun · 10d ago · louisabraham.github.io
- SA
Hello HN! I found that picking out plausible but diverse skin tones for my digital art and game development projects was kind of difficult, and I got curious about if there was a way to define a color space that made it easy. I've built a color picker and procedural generation algorithm based on the space as well as a bunch of other fun js features and demos throughout the page that use the equations. If you find it interesting, I have lots of explanations of how I built it and what properties the space has. The methodology might be a bit shaky, but hopefully the result is as helpful for…
Life & fun · Aug 2026 · toneyalexander.github.io
I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15). The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device. The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.
AI · 17d ago · simedw.com