Highest score on Humanity's Last Exam before Dec 31, 2027?
74.33%median
55%100%
Median 74.33%Middle half 63.16% – 84.78%
Install
Highest score on Humanity's Last Exam before Dec 31, 2026?
The odds
59.49%median
45%95%
Median 59.49%Middle half 57.18% – 64.15%
How the odds have moved
Median: 59.36% on Oct 20, 59.49% now.
The median the market implied over time and its middle half (the band), read from every strike's prices on Kalshi.
Every outcome
| Outcome | Chance | Day | Bid | Ask | Pays | Contracts |
|---|---|---|---|---|---|---|
| At least 50% | 99% | 97¢ | $1.02 | 3.1K | ||
| At least 55% | 99% | 97¢ | $1.02 | 11.7K | ||
| At least 60% | 45% | 42¢ | 47¢ | $2.13 | 15.5K | |
| At least 65% | 21% | ▼ 3 | 19¢ | 23¢ | $4.35 | 10.7K |
| At least 70% | 15% | ▼ 2 | 13¢ | 17¢ | $5.88 | 7.3K |
| At least 75% | 10% | ▼ 1 | 9¢ | 11¢ | $9.09 | 1.9K |
| At least 80% | 5% | ▼ 3 | 4¢ | 5¢ | $20.00 | 2.1K |
| At least 85% | 5% | ▼ 1 | 2¢ | 8¢ | $12.50 | 865 |
| At least 90% | 5% | 3¢ | 7¢ | $14.29 | 1.3K |
"Pays" is what a $1 stake returns if that outcome happens, at the price shown (before any fees). Bid and ask are the best prices to sell and buy "Yes".
The news: Humanity's Last Exam
Search ›-
DeepSeek V4.1 Flash Just Made Frontier Models Obsolete for AgentsDeepSeek just dropped V4.1 Flash 1017 — a minor version bump with major implications. This is their fast/cheap model, and it's beating models that cost 10x more on the benchmarks that matter most for production AI. A Flash model. Beating… -
Kimi K3.1 Pro launches at 4th on the Intelligence Index and will lead open weights modelsMoonshot's new reasoning model gains 11 points over Kimi K3 (max, 44). It trails only Claude Opus 5.5 (max with fallback, 58), Claude Sonnet 5.5 (max with fallback, 56), and Muse Wonder Full (max, 56) on the Intelligence Index. Once the… -
Claude Fable 5.1 Review: The Benchmarks Went Up, but the Cache Price Cut Is the Real ReleaseWhen Anthropic shipped Claude Fable 5 in June, I wrote that the price doubling was the headline, not the benchmarks. Ten dollars per million input tokens and fifty per million output is twice Opus 5, and I said at the time that the… -
# BrowseComp explained: what it measures and how search APIs compareBrowseComp is OpenAI's benchmark for whether an AI agent can track down hard-to-find facts on the open web, and it has become the main stress test for search APIs that feed agents. This guide covers how the questions were built, how… -
An 8B Agent That Thinks Without Writing Tokens Beats GPT-5.6 Luna on GAIA. Not on HLEAn 8-billion-parameter agent does part of its thinking without writing a single token, and on the GAIA agent benchmark it scores 43.7 to GPT-5.6 Luna's 37.9. Those numbers are the authors' own, reported in Table 1 of the Hybrid Reasoning… - DeepSeek Releases V4.1-Flash-1017 With Faster Performance and Improved AI ReasoningChinese artificial intelligence company DeepSeek has released DeepSeek-V4.1-Flash-1017, an updated open-weight AI model that the company says delivers stronger reasoning, coding, AI agent, and visual capabilities while completing tasks…
-
DeepSeek Releases V4.1-Flash-1017 Update, Says It's Faster and LeanerBriefs Finance helps everyday people get better with money and invest with confidence, from getting out of debt all the way to investing like an expert. It starts with Market Briefs, a free five-minute daily newsletter, and grows into… - deepseek-v4.1-flash-1017Which id to call These are the upstream provider rates. Pay as you go adds 5%, or 0% if you bring your own keys, and there is no per-request fee. Prompt caching and routing change what you pay against these rates, not the rates themselves…
-
Nemotron 3.5 Bolt 74B vs Qwen3.8-Max: The Widest Gap on the BoardIt cannot. What makes the pair worth writing about is that the two models win on entirely different rows, and that NVIDIA's own benchmark table compares Bolt to a Qwen model that is not this one. One is a file. One is a service. The… -
Nemotron 3.5 Bolt 74B vs Nemotron 3 Ultra: 1.5 Index Points, Seven Times the ParametersThe two models are not interchangeable, and the places where they differ are more instructive than the headline score. What the compression preserved and what it discarded are both visible in the published numbers, and they are not the… -
Ornith 1.5 9B - AI Model Library | Build on Vast.aiLightweight 9B reasoning model from Ornith, built on Qwen3.5 and trained by self-improving reinforcement learning for agentic coding and tool use Ornith-1.5-9B is the smallest member of the Ornith-1.5 family from Ornith AI, a 9B dense… -
Artificial Analysis Decision Index v1.0The Artificial Analysis Decision Index benchmarks decision models, a new category of models that answer structured decisions in a single call with a probability for each option. Version 1.0 combines four evaluations across classification…
Related markets
All ai ›Which company has the best AI model end of October?
85%Google
- Google85%▼ 4
- Anthropic13%▲ 5
- OpenAI1%
18 more
- Meta<1%
- Moonshot<1%
- Alibaba<1%
- Baidu<1%
- Z.ai<1%
- Xiaomi<1%
- SpaceXAI<1%
- DeepSeek<1%
- Amazon<1%
- ByteDance<1%
- Nvidia<1%
- Thinky<1%
- MiniMax<1%
- Mistral<1%
- Meituan<1%
- Tencent<1%
- StepFun<1%
- Microsoft<1%
Next Fable Model (5.2+) released on...?
69%No release by October 31
- No release by October 3169%▲ 5
- October 2716%▼ 3
- October 286%▲ 2
11 more
- October 264%
- October 294%▼ 1
- October 301%▼ 1
- October 201%▼ 12
- October 31<1%▼ 1
- October 19<1%▼ 2
- October 21<1%▼ 5
- October 22<1%▼ 14
- October 23<1%▼ 4
- October 24<1%
- October 25<1%▼ 1
Which company has best AI model end of 2026?
54%Anthropic
- Anthropic54%▲ 4
- Google33%▼ 6
- OpenAI6%
12 more
- Meta1%
- xAI1%
- Alibaba<1%
- Moonshot<1%
- DeepSeek<1%
- Z.ai<1%
- Mistral<1%
- Microsoft<1%
- ByteDance<1%
- Meituan<1%
- Amazon<1%
- Baidu<1%
Next Claude Haiku Model (4.6+): Text Arena Debut?
1,410median
1,3801,480
Median 1,410Middle half 1,405 – 1,415
Next Fable Model (5.2+) released by...?
Nov 19, 2026even chance by
- October 21<1%▼ 12
- October 22<1%▼ 14
- October 24<1%▼ 1
2 more
- October 3131%▼ 4
- December 3194%▼ 1
Odds are traders' prices on Kalshi, Polymarket, ForecastEx and Crypto.com, not forecasts by 4Search. Prices read from the venues 2 min ago. Prices may be delayed. Money traded on Kalshi and ForecastEx is estimated from contracts. Not betting or investment advice.
Prediction-market prices are what traders pay for each outcome on Kalshi, Polymarket, ForecastEx and Crypto.com, read as the market's implied chance. They are not forecasts by 4Search, can move fast and can be wrong; they are read every few minutes and may be delayed. Money traded on Kalshi and ForecastEx is estimated from contracts at their price. News comes from the 4Search index; prices of stocks, rates and coins from 4Search Finance. Nothing here is betting, financial or investment advice. Terms of use. How to use Predictions