https://modevc.com/charts/One AI chart a day from Arena leaderboard data, frontier-lab launches and public datasets, read for distribution by Mode VC.en Fri, 16 Oct 2026 13:00:00 GMT https://modevc.com/charts/pre-seed-valuation-dispersion/https://modevc.com/charts/pre-seed-valuation-dispersion/Fri, 16 Oct 2026 13:00:00 GMT Pre-seed valuations have fractured into a bifurcated market. Top-decile founders command $28M caps on distribution strength, while thin model wrappers trade at steep sub-$10M discounts. The top 10% of pre-seed AI startups command 3.4x higher valuations ($28.0M vs $8.2M) than bottom-quartile generalist SaaS.pre seed valuation 2026ai startup valuation dispersionventure capital distribution moats https://modevc.com/charts/ai-chatbot-adoption-by-age-demographics/https://modevc.com/charts/ai-chatbot-adoption-by-age-demographics/Thu, 15 Oct 2026 13:00:00 GMT Consumer AI habits are generational. Young demographics show over 60% adoption, while enterprise-grade challengers like Claude remain constrained by niche distribution. 61% of Americans aged 18–29 use ChatGPT; Claude reaches 6% of all adults.ai chatbot adoption by agechatgpt user demographicsclaude vs chatgpt users https://modevc.com/charts/open-weights-vs-proprietary-arena-gap/https://modevc.com/charts/open-weights-vs-proprietary-arena-gap/Wed, 14 Oct 2026 13:00:00 GMT The performance gap between open and closed models has stabilized between 30 and 45 Arena points. For 80% of enterprise tasks, open-weight economics outweigh the marginal frontier lead. Open weights briefly beat every proprietary model in early 2025 (deepseek-r1); today the best open model trails by 42 points.open source vs closed aiopen weights arena leaderboarddeepseek vs proprietary llm https://modevc.com/charts/arena-rating-gap-google-anthropic-openai/https://modevc.com/charts/arena-rating-gap-google-anthropic-openai/Tue, 13 Oct 2026 13:00:00 GMT Benchmark dominance has cycled from OpenAI to Anthropic to Google. The gap between frontier labs is narrow enough that distribution speed decides enterprise adoption. Google leads the Arena again; Anthropic trails by 22 points and OpenAI by 62.arena rating gapllm leaderboard gapgoogle vs anthropic vs openai https://modevc.com/charts/best-ai-models-by-category/https://modevc.com/charts/best-ai-models-by-category/Mon, 12 Oct 2026 13:00:00 GMT Frontier benchmarks are fragmenting by task. Google dominates technical and quantitative categories, while Anthropic retains the lead in expert knowledge and writing. Google holds #1 in 6 of 8 Arena categories; Anthropic in 2. OpenAI has no category in the top 15.ai model comparisonllm benchmark comparison https://modevc.com/charts/ai-inference-cost-comparison/https://modevc.com/charts/ai-inference-cost-comparison/Sun, 11 Oct 2026 13:00:00 GMT Flagship intelligence is decoupling from flagship pricing. Open and low-cost models deliver frontier reasoning at a fraction of the token cost. Qwen3.8-Max rates within 30 Arena points of Claude Opus 5.5 for 3.3× less per output token. DeepSeek V4 Pro is 10.1× cheaper.ai inference cost comparisonllm pricing comparison https://modevc.com/charts/ai-in-the-workplace-statistics/https://modevc.com/charts/ai-in-the-workplace-statistics/Sat, 10 Oct 2026 13:00:00 GMT Workers who use AI mostly report saving time. Fewer than half say their work got better, and only 28% of employees use AI weekly. Among US employees who use AI, 63% say it makes them faster but only 46% say it raises quality. Just 28% of all employees use AI weekly.ai in the workplace statisticsai at work surveygallup ai workplace https://modevc.com/charts/ai-coding-assistant-statistics/https://modevc.com/charts/ai-coding-assistant-statistics/Fri, 09 Oct 2026 13:00:00 GMT 66% of developers use an AI coding assistant, but only 48% trust its output even when they can check it. The open position is the check, not the assistant. 66% of developers use AI coding assistants, but only 48% trust AI output they can validate, a gap of 18 points. Agents reach 26%.ai coding assistant statisticsdevelopers trust ai codestack overflow developer survey 2026 ai https://modevc.com/charts/ai-adoption-rate-us-businesses/https://modevc.com/charts/ai-adoption-rate-us-businesses/Thu, 08 Oct 2026 13:00:00 GMT 23.8% of US businesses now use AI to make what they sell. The other 76.2% will not shop for a model; they will buy through a channel. 23.8% of US firms used AI in the latest two-week survey, up from 17.3% 10 months earlier. 76.2% still did not.ai adoption by industryai adoption rateai adoption statistics 2026 https://modevc.com/charts/llm-api-pricing-history/https://modevc.com/charts/llm-api-pricing-history/Wed, 07 Oct 2026 13:00:00 GMT Frontier input prices fell 56% a year at OpenAI and 40% a year at Anthropic's top tier. Every competitor gets the same price cut. OpenAI's headline model fell from $30 to $2 per million input tokens since Jun 2023: 93% in total, 56% a year.llm pricing comparisonai inference cost comparisonllm price per million tokens https://modevc.com/charts/open-source-vs-closed-ai-models-gap/https://modevc.com/charts/open-source-vs-closed-ai-models-gap/Tue, 06 Oct 2026 13:00:00 GMT The best open-weight model would win 44% of blind votes against the best closed one. The model is becoming a commodity input. The best open model trails the best closed model by 42 Arena points. In a blind vote it would still win 44% of the time, up from 31% at the widest gap.llm benchmark comparisonopen source vs closed source llmbest open source llm https://modevc.com/charts/best-ai-model-vs-most-popular-ai-chatbot/https://modevc.com/charts/best-ai-model-vs-most-popular-ai-chatbot/Mon, 05 Oct 2026 13:00:00 GMT Anthropic builds more top-50 models than any other lab. ChatGPT still reaches about 7 times as many Americans as Claude. Anthropic builds 14 of the Arena's top 50 models and OpenAI builds 5. But 44% of US adults use ChatGPT and 6% use Claude.ai model comparisonmost popular ai chatbotchatgpt vs claude usage https://modevc.com/charts/how-long-number-one-ai-model-stays-on-top/https://modevc.com/charts/how-long-number-one-ai-model-stays-on-top/Sun, 04 Oct 2026 13:00:00 GMT A frontier model now holds Arena's #1 spot for about 3 weeks. A lab that keeps shipping can hold it for months. Since Jan 2025, Arena's #1 spot has changed hands 16 times. Half of the finished reigns lasted 23 days or less.llm leaderboardllm leaderboard historybest ai model right now https://modevc.com/charts/ai-model-launch-strategy/https://modevc.com/charts/ai-model-launch-strategy/Sat, 03 Oct 2026 13:00:00 GMT Most new models now launch to developers first: 9 of 13 opened an API on day one. Only 2 went straight into a consumer app, and both came from labs that own one. Of 13 model launches from 23 Sept 2026 to 3 Oct 2026, 9 opened an API on day one, 7 of them with a price. Only 2 reached a consumer app the same day.ai model launch strategyapi first vs app firsthow ai labs launch models https://modevc.com/charts/openai-python-package-downloads-pypi-growth/https://modevc.com/charts/openai-python-package-downloads-pypi-growth/Fri, 02 Oct 2026 13:00:00 GMT Since PyPI's counting change, the openai package grew 14%, exactly the median of 11 AI packages. anthropic grew 25%. In a market growing this fast, keeping pace is not traction. Between 31 Aug–6 Sep and 25 Sep–1 Oct, anthropic's PyPI downloads grew 25% and openai's 14%, the median of 11 AI packages.openai python package downloadsanthropic python sdk downloadspypi download stats ai https://modevc.com/charts/voice-ai-models-2026/https://modevc.com/charts/voice-ai-models-2026/Thu, 01 Oct 2026 13:00:00 GMT Audio is now a regular slot in the release calendar: 6 labs shipped voice or speech models since July. When everyone sells a voice, the value moves to where the calls happen. Audio and speech made 11 of 59 named model launches since 6 Jul 2026, from 6 labs. In the lab-post sweep since 23 Sep it is 6 of 15.text to speech ai models 2026voice ai modelsspeech to text ai models https://modevc.com/charts/gemini-leaderboard-ranking-days-at-number-one/https://modevc.com/charts/gemini-leaderboard-ranking-days-at-number-one/Wed, 30 Sep 2026 13:00:00 GMT Google had zero days at #1 on Arena in 2024. Since mid-2025 its models have held the top for 227 days, and Gemini 4 Argon just took it back. Since Jan 2024, Google's models have been #1 on Arena for 227 of 1003 days, third among labs behind OpenAI (461) and Anthropic (252). All of it came from 24 Jun 2025 on.gemini leaderboard rankinggemini 4 argon leaderboardllm leaderboard https://modevc.com/charts/openai-leaderboard-ranking/https://modevc.com/charts/openai-leaderboard-ranking/Tue, 29 Sep 2026 13:00:00 GMT OpenAI has not had the top model on Arena for 458 days, and now ranks 8th among labs. ChatGPT kept growing anyway. OpenAI was the #1 lab on Arena for 95% of 2024. On 25 Sept 2026 it ranks 8th among labs, and its best model, gpt-5.5-high, is #26.openai leaderboard rankingchatgpt ranking llm leaderboardllm leaderboard https://modevc.com/charts/open-weight-models-2026/https://modevc.com/charts/open-weight-models-2026/Mon, 28 Sep 2026 13:00:00 GMT Few launches since July said their weights were public, and none came from the three biggest US labs. Open weights now arrive mostly as specialist parts. Of 51 named model launches since 6 Jul 2026, 22 stated their weights: 18 closed, 4 open. All 16 from OpenAI, Google and Anthropic were closed.open weight models 2026open weight llmopen vs closed ai models https://modevc.com/charts/ai-model-releases-2026/https://modevc.com/charts/ai-model-releases-2026/Sun, 27 Sep 2026 13:00:00 GMT 50 named AI models launched in 12 weeks, about one every 1.7 days. 14 labs shipped them; none of them can count on a quiet month. 14 labs launched 50 named AI models from 6 Jul 2026 to 27 Sept 2026. Alibaba's Qwen team led with 7; Chinese labs made 19.ai model releases 2026new ai modelsai model release timeline https://modevc.com/charts/generative-ai-usage-statistics/https://modevc.com/charts/generative-ai-usage-statistics/Sat, 26 Sep 2026 13:00:00 GMT More than half of working-age Americans now use generative AI outside work. The rest will not go looking for it; it will arrive as a default. 54.8% of Americans aged 18–64 used generative AI outside work in Q2 2026, up from 36.0% in Q3 2024. The share passed half in Q1 2026.generative ai usage statisticshow many people use aigenerative ai adoption rate https://modevc.com/charts/chatbot-arena-votes-by-ai-lab/https://modevc.com/charts/chatbot-arena-votes-by-ai-lab/Fri, 25 Sep 2026 13:00:00 GMT OpenAI models have drawn the most Arena votes of any lab, yet hold none of the top 20 seats. Votes measure presence over time, not today's rank. OpenAI's models hold 17.9% of all 12.3M Arena votes and none of the top 20 seats. Anthropic's hold 17.0% of votes and 9 seats.chatbot arena voteslmarena voteslmarena leaderboard https://modevc.com/charts/chatgpt-vs-gemini-vs-meta-ai-users/https://modevc.com/charts/chatgpt-vs-gemini-vs-meta-ai-users/Thu, 24 Sep 2026 13:00:00 GMT All three now claim close to a billion users. Gemini and Meta AI got there by riding products people already use. ChatGPT reports 900M weekly users (Feb 2026). Gemini reports 950M monthly (Jul 2026), 2.4× its May 2025 figure. Meta AI reports 1B monthly (May 2025).chatgpt usersgemini usersmeta ai usershow many people use chatgpt https://modevc.com/charts/ai-productivity-statistics/https://modevc.com/charts/ai-productivity-statistics/Wed, 23 Sep 2026 13:00:00 GMT Claude users say AI makes them faster and lets them take on more. Only 27% say it saves money: AI sells as more output, not lower cost. 86% of surveyed Claude users say AI makes them faster and 82% say it widens what they can do. Only 27% report cost savings.ai productivity statisticsdoes ai make you more productiveai productivity survey 2026 https://modevc.com/charts/ollama-downloads-local-llm-adoption/https://modevc.com/charts/ollama-downloads-local-llm-adoption/Tue, 22 Sep 2026 13:00:00 GMT Ollama's JavaScript client grew 2.1×, but the cloud SDKs grew 8.7×. Local AI is growing; it is not where most products are built. On npm, Ollama's client fell from 27 to 6.6 downloads per 1,000 cloud lab SDK downloads in 46 weeks. On PyPI, after its counting change, it is 21.5.ollama downloadslocal llm adoptionollama vs openai api https://modevc.com/charts/langchain-alternatives-downloads-share/https://modevc.com/charts/langchain-alternatives-downloads-share/Mon, 21 Sep 2026 13:00:00 GMT LangChain's JavaScript core grew 3.2× in 46 weeks, yet its share of framework installs fell from 32% to 19%. An early default can grow and still fade. LangChain's share of weekly npm framework downloads fell from 32% to 19% in 46 weeks, while its own downloads grew 3.2×.langchain alternativeslangchain downloadslangchain vs vercel ai sdk https://modevc.com/charts/mcp-adoption-model-context-protocol-downloads/https://modevc.com/charts/mcp-adoption-model-context-protocol-downloads/Sun, 20 Sep 2026 13:00:00 GMT The MCP SDK is the most downloaded AI package on npm in our set, 1.3× the OpenAI SDK. A protocol spreads through other people's products. The MCP SDK had 52.3M npm downloads in the week of 24 Aug, 1.3× the OpenAI SDK and more than any of the 12 other AI packages we track.mcp adoptionmodel context protocol downloadsmcp sdk npm https://modevc.com/charts/vercel-ai-sdk-downloads-vs-lab-sdks/https://modevc.com/charts/vercel-ai-sdk-downloads-vs-lab-sdks/Sat, 19 Sep 2026 13:00:00 GMT Vercel's AI SDK grew 6.5× in under a year, but the lab SDKs grew 8.7×. The layer that makes models swappable did not capture the growth. The Vercel AI SDK fell from 31 to 23 weekly downloads per 100 downloads of the OpenAI, Anthropic and Google SDKs, after a peak of 40.vercel ai sdk downloadsvercel ai sdk vs openai sdkai sdk npm https://modevc.com/charts/anthropic-sdk-vs-openai-sdk-npm-downloads/https://modevc.com/charts/anthropic-sdk-vs-openai-sdk-npm-downloads/Fri, 18 Sep 2026 13:00:00 GMT In 46 weeks the Anthropic SDK went from 34% to 97% of the OpenAI SDK's npm downloads. Developer distribution can be won late, and lost fast. Anthropic's SDK went from 34% to 97% of the OpenAI SDK's weekly npm downloads in 46 weeks. It passed OpenAI in one week (25 May).anthropic sdk downloadsopenai npm downloadsanthropic vs openai api usage https://modevc.com/charts/llm-rankings-by-category-specialists/https://modevc.com/charts/llm-rankings-by-category-specialists/Thu, 17 Sep 2026 13:00:00 GMT 22 of the Arena's top 60 models rank at least 10 places higher in one category than overall. The overall list hides specialists. claude-opus-4-5-20251101-high-32k is #49 overall but #14 in coding. gpt-5.6-terra-xhigh is #52 overall but #24 in expert prompts.llm leaderboardllm rankings by categorybest llm for hard prompts https://modevc.com/charts/best-ai-for-creative-writing/https://modevc.com/charts/best-ai-for-creative-writing/Wed, 16 Sep 2026 13:00:00 GMT claude-fable-5.1-max leads Arena creative writing, with 9 more models inside its interval. Some models that rank high overall fall a long way here. Anthropic has 7 of the top 10 creative writing models and Google 3. Meta's muse-spark-1.2 (xHigh) is #11 overall but #31 in creative writing.best ai for creative writingbest ai for writing storiesai model comparison https://modevc.com/charts/best-ai-for-math/https://modevc.com/charts/best-ai-for-math/Tue, 15 Sep 2026 13:00:00 GMT claude-opus-5-max tops Arena math, but 14 models from 5 labs are statistically tied with it. 4 of them are lighter "Flash" models. 14 models are statistically tied for #1 in Arena math, spanning 49 points. 4 are Flash models from Google and Z.ai.best ai for mathbest llm for mathllm benchmark https://modevc.com/charts/llm-leaderboard-top-10-gap/https://modevc.com/charts/llm-leaderboard-top-10-gap/Mon, 14 Sep 2026 13:00:00 GMT The #10 model on the Arena board would now win 47% of blind votes against #1. When the top ten are this close, quality stops deciding the sale. The gap between Arena's #1 and #10 models fell from 148 points in Feb 2024 to 18. #10 now wins 47% of head-to-head votes against #1, up from 30%.llm leaderboard top 10llm leaderboardllm benchmark https://modevc.com/charts/chinese-ai-models-vs-us-models/https://modevc.com/charts/chinese-ai-models-vs-us-models/Sun, 13 Sep 2026 13:00:00 GMT The best Chinese model trails the best US model by 27 Arena points, down from 156 in Jan 2024. A year ago it was 27. qwen3.8-max trails claude-fable-5.1-max by 27 points. Over the past year the gap has stayed between 25 and 46; it was 156 in Jan 2024.chinese ai modelsdeepseek vs chatgptai model comparison https://modevc.com/charts/claude-vs-chatgpt-leaderboard/https://modevc.com/charts/claude-vs-chatgpt-leaderboard/Sat, 12 Sep 2026 13:00:00 GMT Claude's best model rates 40 points above OpenAI's best on the Arena board. The lead has changed hands 15 times since 2024, usually by a few points. On 11 Sept 2026, Claude's best model rates 40 points above OpenAI's and 16 above Gemini's. Against OpenAI's best it would win about 56% of blind votes.claude vs chatgptclaude vs chatgpt vs geminiai model comparison https://modevc.com/charts/anthropic-vs-openai-vs-google-arena-top-20/https://modevc.com/charts/anthropic-vs-openai-vs-google-arena-top-20/Fri, 11 Sep 2026 13:00:00 GMT Two labs now hold 15 of the Arena's top 20 spots. OpenAI, which held the most in Apr 2025, holds none. Anthropic holds 8 of the Arena top 20, Google 7 and OpenAI 0. In Apr 2025 OpenAI held 9 and Anthropic 0.anthropic vs openai vs googlellm leaderboardopenai vs anthropic https://modevc.com/charts/how-many-ai-models-are-there/https://modevc.com/charts/how-many-ai-models-are-there/Thu, 10 Sep 2026 13:00:00 GMT The Arena text board now rates 399 models from 52 labs. Only 5 labs have a model in its top 20. Supply keeps growing; the top keeps narrowing. Arena rates 399 text models from 52 labs, up from 10 models in May 2023. Yet only 5 labs hold its top 20 seats, down from 15 in Dec 2023.how many ai models are therehow many llms are thereai model comparison https://modevc.com/charts/best-llm-for-coding/https://modevc.com/charts/best-llm-for-coding/Wed, 09 Sep 2026 13:00:00 GMT Anthropic has had the top coding model on every Arena snapshot for 9 months. Two lighter models from other labs are now inside its margin of error. claude-opus-4-6-high leads Arena coding at 1535. Anthropic has held #1 for 280 days, but Z.ai and Google sit within the interval, 19 points back.best llm for codingbest ai for codingcoding llm leaderboard