AI Tools News 2026: Major Launches, Updates, Pricing Changes
TL;DR
Q1 2026 headlines: Granola raised $125M at $1.5B valuation. Cursor hit 1.0. Anthropic released Claude Opus 4.6. Wispr Flow expanded to Windows. Q2 2026 headlines: ChatGPT briefly cut Pro to $120/mo (was $200), then moved to a two-tier $100/mo (5x) or $200/mo (20x) Pro. OpenAI Codex updates. Cody discontinued its free/pro individual tiers. Replit Agent launched Pro tier. LM Studio announced Enterprise. Mid-July 2026: GPT-5.6 became ChatGPT's default on July 9, xAI shipped Grok 4.5 on July 8, and Fable 5 took over as the top-tier Claude model.
Late May to June 2026: Claude Opus 4.8 shipped, GPT-5.5 Instant became ChatGPT's default model, Google cut AI Plus to $4.99/mo to open a consumer AI price war, and GitHub Copilot moved to usage-based AI Credits.
A chronological roundup of the most significant AI tool announcements of 2026: new launches, model releases, pricing changes, funding rounds, and category shifts. Organized by quarter so you can track the news that actually moved the market, not every hype-cycle vendor blog post.
Stay ahead, get news weekly
Subscribe free →Quick navigation
Q1 2026: January to March
Q1 was dominated by three themes: agentic AI going mainstream, major funding rounds in the productivity category, and frontier model upgrades from Anthropic and OpenAI.
Granola raises $125M Series C at $1.5B valuation
The AI meeting-notes category got its defining headline of the year when Granola closed a $125M Series C at a $1.5B valuation. The round signaled that investors see meeting AI as a durable enterprise category, not a ChatGPT feature. Granola competes directly with Otter.ai, Fathom, and Fellow, and has quickly become the default pick for founders and operators who want elegant, Mac-first meeting notes. See our full Granola review.
Cursor 1.0 goes general availability
Cursor hit its 1.0 milestone in Q1, marking its transition from beta product to stable enterprise tool. The release packaged BugBot for code review, a refreshed Agent mode, and deeper MCP support. Cursor's arc, from a forked VS Code experiment to the dominant AI IDE, is one of the most important stories of the 2025-2026 coding-tools cycle. Compare it head-to-head against Windsurf and GitHub Copilot.
Anthropic releases Claude Opus 4.6
Anthropic released Claude Opus 4.6 in Q1 2026, further extending its lead on reasoning benchmarks and code generation. The release added a 1M-token context window option for enterprise users, the same capacity that powers this very blog post generation workflow at ToolChase. Opus 4.6 also improved tool use (MCP), vision capabilities, and reduced hallucination rates on factual benchmarks. Compare against ChatGPT.
Claude Code goes general availability
Claude Code, Anthropic's CLI-based coding agent, exited research preview and entered general availability. The tool runs autonomous coding sessions from the terminal and has become a favorite of infrastructure and devops engineers who prefer CLI workflows over IDE-based agents. Bundled into Claude Pro at $20/month, it offers exceptional value compared to standalone coding agents like Devin.
Wispr Flow expands to Windows
Mac-first voice interface Wispr Flow launched its Windows version in Q1, dramatically expanding its addressable market. Wispr's core bet is that voice dictation with AI cleanup is 3x faster than typing for most knowledge work, a claim we've tested in our Wispr Flow review.
Q2 2026: Early news (April)
ChatGPT Pro repriced to $120/mo (down from $200) in April
ChatGPT Pro, OpenAI's power-user plan launched at $200/month in late 2024, was cut to $120/mo in April 2026 as it matured into a credible enterprise seat. At the time it included unlimited GPT-5 Pro access, o1 Pro mode reasoning, extended Sora video generation, and Operator agent sessions. (OpenAI has since replaced that flat $120 plan with a two-tier Pro at $100/mo for 5x usage or $200/mo for 20x usage, see the pricing tracker below.) For power researchers and developers, it remains the most expensive single-seat AI subscription short of API budgets, but the math makes sense for teams running deep-research workflows daily.
OpenAI Codex ships updates
OpenAI's Codex CLI, a direct response to Claude Code, shipped significant updates in Q2 including a cloud execution environment and deeper GitHub integration. This cements CLI-based autonomous coding as a durable category, not a one-off feature. See our AI Coding Agents 2026 roundup for context.
Replit Agent adds Pro tier
Replit Agent introduced a Pro tier aimed at serious builders and indie hackers who need more usage than the Core plan delivers. Replit continues to be the best option for non-technical founders who want to ship real apps without writing code.
LM Studio announces Enterprise
LM Studio announced an Enterprise tier for organizations running local LLMs at scale. This is a meaningful milestone, local LLM tools are typically developer-focused and free. LM Studio Enterprise suggests that local AI has graduated from hobbyist curiosity to a real enterprise procurement line, driven by privacy, compliance, and cost pressures at frontier API rates.
May 2026: latest verified changes
These changes were re-verified directly against each vendor's live pricing page in May 2026 (see each tool's "Last verified" date).
Notion AI shifts to credit-based pricing
Notion restructured its AI add-on from a flat per-seat fee to a credit-based model (about $10 per 1,000 monthly AI credits) and adjusted core plans, Plus dropped to $10/mo while Business rose to $20/mo. It mirrors a broader 2026 shift toward usage-metered AI pricing.
Grammarly renames "Premium" to "Pro"
Grammarly retired the "Premium" label in favor of "Pro" ($12/mo), aligning with the now-standard Free / Pro / Enterprise ladder. Features are unchanged, only the tier name moved.
HeyGen adds a Pro tier and raises Business
HeyGen introduced a new Pro plan at $49/mo (between Creator $29 and Business) and raised Business from $89 to $149/mo as it leans into team and enterprise avatar-video workflows.
Perplexity launches a Max tier
Perplexity added a Max tier at $200/mo ($2,000/yr annual), bundling Perplexity Computer credits and Model Council access for heavy research users, its highest consumer plan to date. Pro remains $20/mo.
Late May and June 2026: latest updates
Early summer brought a frontier-model refresh and the first real consumer AI price war. Each item below was verified against the vendor announcement or a primary news source in June 2026.
Anthropic releases Claude Opus 4.8
On May 28, 2026, Anthropic shipped Claude Opus 4.8, building on Opus 4.7 with gains in coding, agentic tasks, vision, and reasoning, plus a new "dynamic workflows" capability in Claude Code for large multi-step jobs. API pricing held steady at $5 per million input tokens and $25 per million output tokens, with up to 90% savings via prompt caching. It is available to Claude Pro, Max, Team, and Enterprise users and across the major clouds. Compare ChatGPT vs Claude.
OpenAI makes GPT-5.5 Instant the default ChatGPT model
OpenAI set GPT-5.5 Instant as the new default model in ChatGPT, with a personalization update reaching the Free and Go tiers on June 9, 2026. OpenAI reported lower hallucination rates in sensitive domains such as law, medicine, and finance, plus stronger math and multimodal reasoning. Plus stays $20/mo and Pro is now $100/mo (5x usage) or $200/mo (20x usage). GPT-5.6 has since replaced it as the default (see the mid-July model wave).
Google opens an AI subscription price war
On June 9, Google cut Google AI Plus from $7.99 to $4.99 per month and doubled its included storage to 400GB, a move TechCrunch called the opening shot in the consumer AI price wars. It followed Google I/O on May 19, where Google added a $100/mo AI Ultra tier and cut the top AI Ultra plan from $250 to $200 (both include 20TB of storage).
GitHub Copilot moves to usage-based AI Credits
From June 1, every GitHub Copilot plan bills on consumed GitHub AI Credits, with a monthly allotment per tier and an optional overage budget. Code completions and Next Edit Suggestions stay unlimited and do not draw credits, while other features are metered by token usage. It is the clearest sign yet that even flat-rate coding tools are shifting to metered AI pricing.
xAI brings Grok 4.3 to Amazon Bedrock
On June 15, xAI's Grok 4.3 arrived on Amazon Bedrock, the first xAI model on the platform and the third major independent lab there alongside Anthropic and OpenAI. AWS describes it as a reasoning-first model with configurable reasoning effort and a 1M-token context window.
Salesforce signs a deal to acquire Contentful
On June 1, Salesforce announced a definitive agreement to acquire headless CMS company Contentful, adding an API-first content layer to Agentforce so AI agents can assemble and deliver personalized content automatically. Terms were not disclosed, and the deal is expected to close in Salesforce's fiscal 2027. It is another sign that agent platforms are buying up the content and data layers around them.
July 2026: the monthly pricing sweep
On July 4, 2026 ToolChase re-verified pricing and versions across 25 flagship AI tools against vendor pages. These are the changes that checked out, each reflected across our tool pages and comparisons the same day.
OpenAI splits ChatGPT Pro into two tiers
The flat $120/mo Pro plan is gone: ChatGPT Pro now comes as $100/mo with 5x usage or $200/mo with 20x usage. The budget Go tier quietly rose from $6 to $8/mo, Business moved to $20/user/mo on annual billing ($25 monthly). Five days later the model itself changed too (see the mid-July model wave below). Full analysis in Is ChatGPT Pro worth it.
Claude Team splits into Standard and Premium
Anthropic replaced the flat $30/user Team plan with Team Standard at $25/mo ($20 annual) and Team Premium at $125/mo ($100 annual), both with a 5-seat minimum. Claude Sonnet 5 shipped June 30, joining Opus 4.8 in the lineup. Anthropic also launched Claude Science, a research workbench app bundled into all paid plans (macOS and Linux beta).
Google cuts AI Plus to $4.99 and splits Ultra
Google AI Plus dropped from $7.99 to $4.99/mo (with 400GB storage) and the single $249.99 Ultra plan became two tiers at $99.99/mo and $200/mo. The app now runs on the Gemini 3.5 family. See ChatGPT vs Gemini for how the ladders compare.
GitHub Copilot completes the move to usage-based AI Credits
Since June 1, GitHub Copilot meters chat through monthly AI Credits: Pro $10 (+$15 credits), the new Pro+ $39 (+$70) and Max $100 (+$200), Business $19 and Enterprise $39 per user. Code completions stay unlimited on paid plans. Separately, Microsoft retired the consumer Copilot Pro plan for new subscribers; the consumer path is now Microsoft 365 Premium at $19.99/mo, with the business add-on holding at $30/user/mo.
Runway retires Unlimited, replaces it with Max
Runway stopped selling the $95 Unlimited plan to new subscribers on June 1. Its replacement, Max, keeps the $95/mo price ($76 annual) but is purely credit-based at 9,500 credits/mo with one-month rollover. Existing Unlimited subscribers migrate September 1.
Smaller July moves worth knowing
Midjourney made V8.1 the default model (June 11). Surfer SEO retired its Essential and Scale tiers for a five-tier lineup starting at Discovery $49/mo. Ahrefs added a $29/mo Starter plan and published Enterprise pricing at $1,499/mo. Synthesia raised Starter to $29 and Creator to $89 but now includes a genuinely recurring free plan (10 video minutes per month). Leonardo AI renamed its tiers to Essential, Premium, and Ultimate with prices unchanged. DeepSeek V4 Flash holds at $0.14 per million input tokens ($0.0028 on cache hits). Perplexity Max now lists at $200/mo. Suno doubled its free tier to 10 songs per day.
Mid-July 2026: the model wave
The second week of July brought the biggest model turnover of the summer: three of the four most-used assistants changed their flagship inside five days. Every change below is already reflected on our tool pages and comparisons.
OpenAI makes GPT-5.6 the ChatGPT default (July 9)
ChatGPT now runs GPT-5.6 as the default model on every tier, including free. The release ships in Sol, Terra, and Luna tiers and replaces GPT-5.5 across the product. Plan-by-plan value is broken down in Is ChatGPT Pro worth it.
xAI ships Grok 4.5 (July 8)
Grok 4.5 landed one day ahead of OpenAI's release and is xAI's strongest model yet for complex reasoning. It is available through X Premium ($8/mo), Premium+ ($40/mo), and the standalone SuperGrok plan ($30/mo), and the platform keeps its 2-million-token context window, the largest among frontier models.
Anthropic's Fable 5 tops the Claude lineup
Claude now offers Fable 5 as its top-tier model, sitting above Opus 4.8 and the Sonnet 5 release from June 30. Free users stay on Sonnet 5; Pro ($20/mo) and Max (from $100/mo) unlock the flagship models. See how the lineup compares in ChatGPT vs Claude.
Pricing changes tracker
- Notion: Plus $12→$10, Business $18→$20; AI add-on moved to credit-based (~$10 per 1,000 credits). Verified May 2026.
- Grammarly: "Premium" plan renamed to "Pro" ($12/mo). Verified May 2026.
- HeyGen: New Pro tier at $49/mo; Business raised $89→$149. Verified May 2026.
- Perplexity: New Max tier at $200/mo (annual). Verified May 2026.
- Google AI Plus: Cut from $7.99 to $4.99/mo; included storage doubled to 400GB. Verified June 2026.
- Google AI Ultra: New $100/mo tier added; top Ultra plan cut $250→$200/mo (20TB storage). Verified June 2026.
- GitHub Copilot: All plans moved to usage-based AI Credits; code completions stay unlimited. Verified June 2026.
- Cody (Sourcegraph): Discontinued Free and Pro individual tiers in 2025-2026, consolidating around the Enterprise Starter plan at $19/user/month. A significant shift for solo developers who relied on Cody's generous free limits.
- Claude: Added Claude Max tier at $100/month for 5x-20x higher usage limits. Claude Pro remains at $20/month.
- ChatGPT: ChatGPT Go at $8/month continues to expand to more countries, offering a lower entry point than Plus at $20. Pro moved from a flat $120/mo to a two-tier $100/mo (5x usage) or $200/mo (20x usage). Verified July 2026.
- Gemini: Gemini Plus at $4.99/month continues to undercut ChatGPT Go slightly and includes 2TB Google One storage.
- Replit Agent: Added Pro tier above the $15/month Core plan.
- LM Studio: Remains free for individuals; added Enterprise tier for orgs.
New launches to watch
Beyond the headline stories, Q1 and early Q2 2026 brought a steady stream of meaningful launches. Notable tools we've tested and added to our directory this year include Lovable (full-stack app builder), Bolt.new (in-browser full-stack agent), v0 by Vercel (UI-first code gen), and Devin (autonomous software engineer). The build-an-app-by-describing-it category has quickly become the most-funded AI subcategory of 2026, with each tool pursuing a slightly different positioning.
Funding rounds that matter
Funding announcements are noisy, but the ones that tend to matter for tool buyers are rounds that (a) extend runway past the current AI capex cycle, (b) signal category leadership, or (c) lock in defensibility around data or integrations. Granola's $125M at $1.5B fits all three. Other 2026 rounds worth noting include continued momentum at Anthropic (OpenAI-comparable valuations), deeper raises for Cursor, Perplexity, and Harvey, and significant Series B rounds across the agentic and AI-voice categories.
Our rule of thumb: if a tool you rely on raises a clean Series B or later at a reasonable valuation, it's probably safe to keep investing your workflows in. If a tool can't raise after 18 months in market, start planning a migration, AI tool consolidation will accelerate throughout 2026.
How we track this news
We track vendor announcements, verify pricing changes directly on official pricing pages, and cross-check funding claims against primary sources. We don't republish press releases uncritically, and we don't take placement payments for inclusion in this roundup. See our methodology for details on how we select and verify news items.
Related resources
FAQ
What was the biggest AI tool story of early 2026?
The GPT-5 launch in January 2026 remains the defining event, it reset the quality bar for general-purpose AI and triggered price cuts across the industry. Other big stories: Anthropic's Claude 4 launch, Google's Gemini 2.5 Pro with improved multi-modal capabilities, Microsoft's Copilot Pro pricing restructure, and the first wave of agentic coding tools (Devin, Claude Code, OpenClaw) reaching real users. The pace of change is the highest in 3 years.
Did ChatGPT raise its prices in 2026?
No, it lowered them. ChatGPT Plus remains $20/mo. ChatGPT Pro is now $100/mo (5x usage) or $200/mo (20x usage), the two-tier structure that replaced the flat $120/mo Pro plan it briefly ran in Q2 2026. Business is $20/user/mo on annual billing ($25 monthly). The Go tier sits at $8/mo for price-sensitive users, and the default ChatGPT model is now GPT-5.6. OpenAI is under competitive pressure from Gemini Free (which is very generous) and has no room to raise consumer prices.
Which AI coding tools discontinued free plans in 2026?
None of the major ones. Cursor free tier is still active (2,000 completions/mo). GitHub Copilot Free has been extended. Replit still has free tier. What changed: Replit lowered free-tier compute in Q1 2026. Claude Code is only on paid Pro ($20/mo) with no free tier. Most rumours about discontinued free tiers turned out to be cosmetic changes to credit limits rather than removals.
What new AI tools launched in Q1 2026?
Notable launches: Anthropic's Claude 4 Sonnet and Opus, Google's Gemini 2.5 Pro, Sora 2 with longer clips, Runway Gen-4, Cursor Composer Agent, Claude Code, Perplexity Spaces, Notion AI 3.0, and a wave of specialised agents (Devin improvements, Factory AI, Magic.dev). The pattern: fewer entirely new tools, more dramatic upgrades to existing leaders. The AI tool market is consolidating around the top 20-30 products.
How should I follow AI tool news going forward?
Best sources in 2026: ToolChase's blog and compare hub (what we do), The Rundown AI newsletter, Latent Space (podcast), Hacker News for technical news, Ben's Bites for quick summaries, and Stratechery for strategic analysis. Avoid Twitter/X hype accounts, signal-to-noise is poor. One newsletter plus one aggregator is enough.
Are there any AI tools that shut down or got acquired in 2026?
Yes, consolidation is accelerating. Jasper was rumoured for acquisition in Q1 (no confirmed deal yet). Smaller 'wrapper' startups shuttered after GPT-5 made their niche obsolete. Stability AI underwent another leadership change. Harvey AI raised a huge round and expanded into new verticals. Mistral raised Series C. Most independent AI tool companies are either growing fast or shrinking fast, very few are stable.
Is AI still improving as fast as in 2024?
Yes, arguably faster. Benchmark jumps from GPT-4 to GPT-5 (and from Claude 3 to Claude 4) are comparable to 2023-2024 jumps. What's new in 2026 is the focus on agents (not just chat), longer reasoning (o1, o1-pro) and multi-modal capabilities (video, audio, code). Costs keep dropping, tokens are 5-10x cheaper than in 2023 for comparable quality. If anything, the rate of practical improvement has accelerated.
Will AI pricing keep falling in 2026?
API prices, yes, expected to drop 2-3x further by end of 2026. Consumer subscription prices, probably stable around $20/mo for Pro tiers, with new budget tiers ($5-10/mo) emerging. Free tiers will get more generous due to competitive pressure from Gemini Free. The biggest cost pressure is on specialised tools (Jasper, Devin) that struggle to justify premiums over general-purpose alternatives at $20/mo.
What should I cancel in my AI subscription stack in 2026?
Audit these suspects: any tool you haven't opened in 30 days, tools with features now replaced by ChatGPT or Claude, specialised writing tools if you already pay for ChatGPT Plus ($20/mo), and 'AI-powered' tools that just wrap OpenAI API (you can use the source directly). Keep: one general AI ($20/mo), one coding AI ($10-20/mo), one domain-specific tool if your job needs it. Under $100/mo total is the sweet spot for most individuals.
Will AI replace jobs faster in 2026?
Yes, for specific roles. Customer support Tier 1, junior copywriting, basic graphic design and some junior coding roles are seeing real headcount reductions (10-25%). Senior roles in the same fields are growing. Professions relatively untouched: nursing, teaching, trades, complex law and medicine. The 2026 pattern: junior-level knowledge work consolidates into fewer people using AI tools; senior judgment work grows.
Are there any AI regulations changing in 2026?
Yes. EU AI Act compliance for high-risk systems becomes enforceable in mid-2026. Several US states (NY, Illinois, California) have new AI-in-hiring rules. The UK is finalising its AI safety framework. China continues tightening AI content rules. For toolbuyers, this means more vendors offering EU-compliance documentation and more disclosure requirements in regulated industries. Expect more 'AI disclosure' notices on consumer products.
What's the next big AI tool category to watch?
Agentic AI, autonomous tools that complete multi-step tasks. Early leaders: Devin, Claude Code, ChatGPT Operator, Google Mariner. 2026 is the first year these are usable for real work, though still unreliable. Also watch: on-device AI (Apple Intelligence, Gemini Nano), AI in spreadsheets beyond Excel, and vertical agents for legal, medical and financial workflows. The next Netflix-of-AI hasn't emerged yet.