LinkedIn Cut Free InMail 87% in Late 2025. Redo the Math.

A September 2026 Glozo analysis flagged that LinkedIn quietly capped Open InMail at under 100 sends per month per account, down from a practical ceiling around 800. That is an 87.5% cut in free outbound capacity, and it lands on top of a Sales Navigator credit ceiling that Leonar re-verified the same month: 50 credits a month, no top-up, no exceptions. If your sourcing plan still treats LinkedIn as the primary channel, the arithmetic no longer works.

What actually changed with the LinkedIn Open InMail limit

LinkedIn compressed the Open InMail limit from a practical ~800/month per account to under 100/month per account in late 2025, according to Glozo's Sept 2026 breakdown. Open InMail is the free-to-send channel to members whose profiles have the "open" flag toggled on, and it was the release valve everyone quietly relied on when paid credits ran dry.

The cut has three practical consequences:

1. Free outbound capacity per account fell roughly 87.5%.
2. Volume gets pushed into paid Credit InMail, which is metered.
3. Templated, AI-personalized outreach now competes for a much smaller share of voice, and reply rates on it are already declining.

Glozo's framing is important: this is not an anti-spam measure. It is demand shaping. LinkedIn engineered the scarcity so that recruiters who used to lean on Open InMail have to buy credits or buy a bigger seat. The mechanism is monetization of attention LinkedIn itself throttled.

87.5%

The Sales Navigator ceiling nobody re-read

Sales Navigator Core and Advanced both cap InMail at 50 credits per month, with a maximum accumulation of 150, and LinkedIn does not sell extras. Leonar re-verified this on Sept 21, 2026, and LinkedIn's own help page (linkedin.com/help/sales-navigator/answer/a101030) confirms the 150-credit accumulation ceiling.

That is the number most sourcing teams have not internalized. If you bought Sales Navigator as your outreach horse because it is cheaper per seat than Recruiter, you are running with a hard 50/month cap and no way to buy more. The credit refunds help - a credit is returned when an InMail is accepted, declined, or answered within 90 days - but they do not lift the ceiling, only the effective cost.

Here is the current 2026 tier table, pulled from Leonar's pricing breakdown:

RPS list price in 2026 is $6,000 to $10,000 per seat per year (about $500 to $833/month), per Glozo's pricing guide. Even at the top of that range, you are buying a metered channel.

The counterintuitive bit

Sales Navigator is now strictly worse than Recruiter Lite for outreach volume per dollar. A Sales Nav Core seat gives you 50 credits. A Recruiter Lite seat gives you 30 credits fresh but a 120-credit rollover ceiling, and the credits behave the same on refunds. Teams keep buying SN because search is unrestricted and because the credit cap does not bite until month two. Re-run the comparison for your actual send volume.

The coverage math is brutal, and this is where teams keep flinching

A single Sales Navigator seat can touch roughly 1 in 1,850 of the Python-fluent US software-engineer population per month. That is the sentence that should be pinned above every sourcing dashboard in Q4 2026.

In Refolk's index of professional profiles, there are 92,860 US software engineers at Software, Senior, or Staff level who list Python. A Sales Nav seat with 50 credits touches 0.054% of that pool in a month. A Recruiter Corporate seat at 150 credits touches 0.16%. If you are hunting a smaller slice, say the 14,573 US machine learning and AI engineers in the same index (top current employers: Meta, Distyl, Berkeley AI Research), one SN seat covers 0.34% per month. That is still one candidate reached for every 291 in the pool.

1 in 1,850

LinkedIn's credit ceilings were designed for a market where the average target pool for a role was measured in low thousands. Modern hiring markets for mainstream stacks are 5x to 50x larger, and the credit math did not move with them. Any strategy that pretends LinkedIn is the primary outreach channel for a mainstream tech stack is arithmetically broken.

This is the exact gap Refolk closes: you describe the person in plain English ("senior Python engineers in Berlin who have shipped async workloads and are not at Zalando"), and I return a ranked shortlist across GitHub, LinkedIn, and the open web so your 50 credits go to the 50 people worth spending them on, not the first 50 the search UI happens to sort to the top.

Where to move the outreach budget

Move the marginal outreach dollar to email, GitHub DMs and issues, and community channels, in that order, and treat LinkedIn as the finishing move, not the opener. The credit math forces this; the reply-rate math confirms it.

Candidates now receive three to ten InMails per week following the same auto-personalized pattern, per Glozo. The fingerprint of machine-generated personalization is recognizable, and templated reply rates have been sliding since 2024. Cutting supply while noise stays high means your reply rate falls faster than your credit count. So:

• Email first. Deliverability is uncapped, cost per send is near zero, and you own the sequence. Verify addresses before sending; sender reputation is now the real budget.
• GitHub second, for engineering hires. Public commits, PR comments, and issue threads give you a topic to open on. Use the notification email on a recent commit for the reach, not the LinkedIn DM.
• Community third. Discords, Slack groups for your niche, conference alumni lists. Warm intros still outperform cold InMail, and they do not consume credits.
• LinkedIn last. Use paid credits on the profiles where email failed and no public contact surface exists.

The Q4 2026 platform limits worth memorizing

Valley's Sept 2026 limits update is the cleanest reference. The hard ceilings for 2026:

• ~100 connection invitations per week (Premium does not raise this).
• ~100 messages/day to existing 1st-degree connections on free accounts, ~150/day on paid.
• Safe automation zones are lower: 20 to 25 invitations/day and 30 to 40 outreach messages/day.

Cross those, and LinkedIn throttles or restricts the account. The Open InMail cut sits on top of these, not instead of them.

The rewrite that makes the math work

Shrink volume, raise pre-filter quality, refund more credits. Glozo's own worked example: at 30% fewer InMails and a 50% higher reply rate, a mid-performing recruiter moves from 250 InMails per placement to roughly 110, and cost per placement drops from $1,400-$3,600 down to $600-$1,100.

That range assumes you can raise reply rate 50%. You cannot do that by writing a cleverer template. You do it by shrinking the target list to people whose public work matches the role, so the opener writes itself and the recipient can tell in one sentence that you actually read something they made. That is a sourcing problem, not a copywriting problem.

Cutting InMail supply while inbox noise stays high means reply rates fall faster than credit counts.

The concrete workflow:

1. Define the role in one paragraph, including two or three signals a real practitioner would recognize (a library, a system, a paper, a conference talk).
2. Pull a candidate pool across GitHub, LinkedIn, and the open web.
3. Rank the pool by signal match, not recency of profile update.
4. Email the top of the ranked list first, GitHub the ones without discoverable emails, and reserve LinkedIn credits for the profiles where both failed.
5. Track refund rate on the LinkedIn credits you do spend. Under the 90-day rule, an accepted, declined, or answered InMail returns the credit. A 40%+ refund rate is the sign your pre-filter is working.

For engineering hiring specifically

If you are running engineering searches, the change hits hardest because engineering pools are the biggest. In Refolk's index, US Python-fluent engineers outnumber their German counterparts about 19 to 1 (92,860 vs 4,891), and Berlin dominates the German slice with Munich second. That kind of asymmetry means your channel mix has to differ by geography: LinkedIn credits go further in a 4,891-person pool than in a 92,860-person one, and GitHub coverage is disproportionately valuable in the US pool.

What to stop doing in Q4 2026

Stop treating InMail volume as the leading indicator of a sourcing team's output. Under the new cap, volume is a proxy for wasted spend, not effort.

The specific practices to retire:

• Ranking sourcers on "InMails sent this week." Rank on reply rate and refund rate instead.
• Buying Sales Navigator as the default seat. Recheck against Recruiter Lite for your actual monthly send volume.
• Running a single-channel LinkedIn sequence for a mainstream stack pool over 20,000 people. The coverage math does not survive it.
• Relying on Open InMail as a fallback when credits run out. It is effectively gone.
• Sending the same AI-personalized opener that three to ten other recruiters sent that week. Recipients pattern-match it in a second.

The Q4 2026 sourcing stack is a pre-filter, an email sender, a GitHub-aware research layer, and LinkedIn as a finishing channel. If you build in that order, the InMail cap stops being a constraint and starts being a feature: it forces the discipline you needed anyway.

FAQ

What is the new LinkedIn Open InMail limit for 2026?

Open InMail is capped at under 100 sends per month per account, down from a practical limit of roughly 800 before the late-2025 change, per Glozo's Sept 2026 analysis. Open InMail is the free channel to members with the "open" flag on their profile, and the cut effectively eliminates it as a bulk outreach tool. Paid Credit InMail is now the primary LinkedIn outreach lane, and Sales Navigator's 50-credits-per-month ceiling has no purchase option to lift it.

Can I still buy more InMail credits on Sales Navigator?

No. Sales Navigator Core and Advanced both include 50 InMail credits per month, cap accumulation at 150, and LinkedIn does not sell add-on credits at any Sales Nav tier. This was reconfirmed on LinkedIn's help page and in Leonar's Sept 21, 2026 pricing update. If you need more paid InMail capacity, the only path is Recruiter Lite (30/mo, 120 ceiling), Recruiter Corporate (150/mo), or a Recruiter Professional Services seat at $6,000 to $10,000 per year.

What are the best alternatives for sourcing outside LinkedIn?

Email, GitHub, and community channels, in that order for most technical roles. Email has no platform-imposed ceiling and near-zero marginal cost, so verified-address outbound is the first move. GitHub gives you a public signal (commits, PRs, issues) and a contactable email on many public profiles, which matters when your target pool is tens of thousands of engineers. Refolk pulls candidates across GitHub, LinkedIn, and the open web from a plain-English brief, so you can rank a pool before you spend any InMail credits at all.

How do I recalculate InMail credit math for a real search?

Divide your monthly credit allotment by the size of your target pool to get monthly coverage percent, then multiply by your reply rate to get expected replies per month. For example, 50 SN credits against 14,573 US ML engineers is 0.34% coverage; at a 10% reply rate that is 5 conversations a month per seat. If the number of conversations you need is higher than that, the search cannot be run on LinkedIn credits alone, and the marginal outreach has to come from email, GitHub, or community.