IMPLICATOR
Implicator's weekly LLM Popularity Meter ranked Gemini above Claude for the first time on May 17, with Gemini at 88 and Claude at 87 on the 100-point enterprise-buyer scorecard. The shift followed Anthropic's May 13 disclosure that programmatic Claude usage will move to a separate metered credit pool on June 15, alongside an Implicator editorial published May 14 that documented the company's per-token API rates and its unpublished subscription token limits. Claude lost two points; the same week's PwC alliance expansion, SAP Sapphire announcement, $200 million Gates Foundation partnership, and Claude for Small Business launch did not offset the drag.
Key Takeaways
- Implicator's weekly LLM Meter ranked Gemini above Claude for the first time on May 17, with Gemini at 88 and Claude at 87 on the 100-point enterprise-buyer scorecard.
- Anthropic moves programmatic Claude usage to a separate metered credit pool on June 15: $20 for Pro, $100 for Max 5x, $200 for Max 20x.
- An Implicator editorial documented that Anthropic publishes per-token API rates but not the token sizes of any subscription's usage limits.
- PwC's 30,000-professional training, SAP Sapphire embed, and a $200 million Gates Foundation partnership did not offset the transparency drag.
AI-generated summary, reviewed by an editor. More on our AI guidelines.
The Agent SDK change
Anthropic's June 15 policy creates monthly programmatic credit pools billed at API-style rates, InfoWorld reported on May 14. Pro subscribers get $20 in credits, Max 5x users $100, and Max 20x users $200. The credits cover the Claude Agent SDK, GitHub Actions integration, the claude -p feature in Claude Code, and third-party frameworks including OpenClaw. Web chat and standard interactive Claude use stay inside the subscription pool.
Senior data scientist Yadesh Salvi wrote on X, per InfoWorld's reporting, that the allocation "won't even last a day of serious work." Advait Patel, a senior site reliability engineer at Broadcom, said enterprises "may find it harder to forecast costs" once usage is tied more directly to token consumption than subscription tiers.
The transparency gap
Anthropic's API pricing page lists Opus 4.7 at $5 and $25 per million input and output tokens. The subscription page lists Pro at $20 a month, Max at $100 or $200, and Team by the seat. Neither the help center nor the Claude Code documentation publishes the token size of any plan's allowance, a May 14 review found. Higher tiers appear in the help center only as multipliers of Pro: Max 5x at five times Pro, Team Premium at 6.25 times.
Get Implicator.ai in your inbox
Strategic AI news from San Francisco. No hype, no "AI will change everything" throat clearing. Just what moved, who won, and why it matters. Daily at 6am PST.
No spam. Unsubscribe anytime.
Anthropic raised Claude Code's weekly usage limits 50 percent on May 13, an increase set to expire July 13. The May 6 post that doubled Claude Code's five-hour session limits and removed a weekday peak-hours throttle also reported Anthropic's lease of SpaceX's Colossus 1 data center, more than 220,000 NVIDIA GPUs and more than 300 megawatts of capacity, per a TechCrunch summary and a writeup from developer Simon Willison cited in the Implicator editorial.
Claude's distribution wins did not offset
PwC and Anthropic announced a joint Center of Excellence on May 14. PwC committed to train and certify 30,000 US professionals on Claude, with Claude Code and Claude Cowork in the broader rollout. Anthropic CEO Dario Amodei said in the announcement that "insurance underwriting that took 10 weeks now takes 10 days" in live deployments. PwC US Senior Partner Paul Griggs said the firm's clients "are looking for ways to apply AI that are secure, responsible, and capable of delivering measurable outcomes."
SAP designated Claude as a primary reasoning and agentic capability on its newly announced SAP Business AI Platform at SAP Sapphire 2026. Anthropic also disclosed a $200 million, four-year Gates Foundation partnership covering global health, education, and economic mobility, and launched Claude for Small Business with a ten-city training tour kicking off in Chicago. The Pentagon's Impact Level 6 and Impact Level 7 classified-network vendor roster, issued May 1, still does not include Anthropic.
Gemini's path to the top
Mars Global CMO Gülen Bengi told Food and Drink Technology in late April that Gemini Enterprise will be the company's primary AI operating system across snacking, petcare, and food and nutrition, citing "One Demand AI" as the rollout's first showcase. Snowflake earlier integrated Gemini 3 natively inside its Cortex AI surface so customer data stays inside the Snowflake perimeter, with BlackLine named as an anchor reference. Google added a $750 million partner-agent innovation fund at Cloud Next '26 to seed forward-deployed work through Accenture and other system integrators.
The Tuesday May 19 keynote at Shoreline Amphitheatre in Mountain View will introduce a new Gemini model. TechTimes preview coverage on May 17 framed the release as landing near GPT-5.5 and short of Anthropic's restricted Claude Mythos preview.
What lands next
Anthropic's next scheduled limit change is July 13, when the 50 percent weekly Claude Code increase is set to expire. Google's May 19 I/O keynote will set the next directional input for both vendors. The Implicator meter's next issue is scheduled for May 24.
Frequently Asked Questions
Why did Gemini pass Claude in the Implicator LLM Meter?
Anthropic's May 13 disclosure that programmatic Claude usage will move to a separate metered credit pool on June 15, plus an Implicator editorial documenting that Anthropic publishes per-token API rates but not the token sizes of its subscription limits, drove Claude down two points. Gemini gained one point on continuing Google Cloud distribution wins, including Snowflake Cortex AI integration and the Mars Gemini Enterprise deployment.
What is the Implicator LLM Popularity Meter?
Implicator's weekly enterprise-buyer scorecard, launched March 30, 2026, ranks Claude, ChatGPT, Gemini, Mistral, Grok, and DeepSeek on a 100-point scale. Scores reflect enterprise factors such as compliance certifications, model quality for professional tasks, reliability, ecosystem maturity, vendor stability, and pricing transparency. The meter publishes new scores every Sunday.
What does Anthropic's June 15 Agent SDK metering change do?
Programmatic Claude usage covers the Claude Agent SDK, GitHub Actions integration, the claude -p feature in Claude Code, and third-party frameworks including OpenClaw. Starting June 15, this draws from a separate monthly credit pool. Pro subscribers get $20, Max 5x users $100, and Max 20x users $200, billed at API-style rates after credits run out. Web chat and interactive Claude use remain inside the subscription pool.
What did the Implicator editorial document about Anthropic's plan limits?
The May 14 editorial documented that Anthropic publishes exact per-token API rates and exact monthly subscription prices but does not publish the token sizes of any plan's usage allowance. A May 14 review of Anthropic's billing pages found no per-plan limit table; higher tiers appear in the help center only as multipliers of Pro, with no published denominator.
When does the Implicator LLM Meter publish next?
The next Implicator LLM Meter issue is scheduled for May 24. Google's I/O 2026 keynote on May 19 and Anthropic's July 13 weekly-limit expiration are the next two scheduled inputs likely to move scores. The meter publishes weekly on Sundays.
AI-generated summary, reviewed by an editor. More on our AI guidelines.
Marcus Schuler
Editor-in-Chief and founder of Implicator.ai. Former ARD correspondent and senior broadcast journalist with 10+ years covering tech. Writes daily briefings on policy and market developments. Based in San Francisco. E-mail: [email protected]
