A Comparative Timeline of Solr, Elasticsearch & OpenSearch
2025 marked a turning point for open source search platforms — with major updates to Solr, Elasticsearch, and OpenSearch redefining how vector search, LLM integration, and enterprise scalability are evolving.
Whether you’re running Apache Solr, Elasticsearch, or OpenSearch, the release cadence, feature focus, and operational maturity have all advanced significantly this year. Below is a month-by-month breakdown of key developments — followed by commentary on what they mean for buyers, architects, and engineers.
In this 2025 Open Source Search Review, we break down every major release and announcement from Solr, Elasticsearch, and OpenSearch — month by month — highlighting how each platform evolved in areas like vector search, LLM integration, and enterprise scalability.
Here’s how the year unfolded across the open source search ecosystem:
Monthly Highlights Table
| Month | Platform | Key Announcement / Release | Why it Matters |
|---|---|---|---|
|
Jan |
Solr |
9.8.0 GA (23 Jan) — Cross-DC Replication graduates from sandbox; memAllowed thread-memory cap added. Source: Apache Solr Release Notes |
Content |
|
Feb |
OpenSearch |
2.19.0 released (11 Feb) with workload management, query insights & template queries. Source: OpenSearch Documentation |
Builds out ML/observability capabilities, making OpenSearch more than “just search”. |
|
Feb |
Elasticsearch |
9.0.0 beta1 (11 Feb) announced. Source: Elasticsearch Release Notes |
Signals the next major version of Elasticsearch is in motion – good time to start planning. |
|
Mar |
Elasticsearch |
9.0.0 RC1 released (20 Mar) ahead of GA. Source: GitClear |
Stabilization phase means early adopters can start pilot plans. |
|
Apr |
Elasticsearch |
9.0.0 GA released (8 Apr) with major vector/quantization and Lucene-10 support. Source: GitClear |
Major version bump – important for compatibility, performance, feature planning. |
|
May |
OpenSearch |
3.0.0 GA released — first major 3.x release (Lucene 10 + JVM 21 + vector/ hybrid search improvements, better performance) Source: OpenSearch |
Major architecture update (Lucene 10, JDK 21) that boosts performance and sets the foundation for future AI and vector search features. |
|
June |
OpenSearch |
3.1 release, with enhancements for vector-driven / AI-driven workloads in cloud / managed environments. Supported on AWS later in September. Source: OpenSearch |
Stabilizes the 3.x line and improves support for vector, neural, and AI-driven workloads – safer for early adopters. |
|
Jul |
Solr |
9.9.0 GA (23 Jul) — adds indexing time text→vector support (via external LLM services), improved fuzzy/rank queries. Source: Apache Solr Release Notes |
Solr moves more directly into the vector/LLM-augmented search world. |
|
Jul |
Elasticsearch |
9.1.0 GA (23 Jul) released; also noted security/3rd-party ecosystem updates around 9.x (e.g., Search Guard FLX support). Source: GitHub |
Shows Elasticsearch 9.x is maturing, ecosystem catching up with security/ops support. |
|
Aug-Sep |
Elasticsearch |
Continued 9.0.x hardening (e.g., 9.0.4 July; 9.0.8 Oct) and 8.19.x maintenance. Source: Elastic |
Upgrade paths remain active; dual-train (8.19 & 9.x) stability matters. |
|
Oct |
Elasticsearch |
9.2.0 GA (approx 23 Oct) and aligned client/SDK updates (e.g., Python client, Java) noted. Source: GitHub |
Illustrates fast minor version cadence – ensures new features keep flowing. |
|
Oct |
OpenSearch |
3.3.0 GA – newest major release in 2025, part of regular ~8-week minor release cadence. Source: OpenSearch 3.3.2 patch is also released later in the month. |
Most stable 3.x release to date with broad performance tuning – recommended baseline for teams planning migration from 2.x. 3.3.2 patch includes critical fixes to vector / k-NN, ML, and security plug-ins – strengthens production readiness for AI search workloads |
Note: Some months had multiple updates (e.g., Feb for both OpenSearch & Elasticsearch). The table picks out the most salient one per platform per month for readability. The frequency of how often a platform is listed is also not indicative of more feature advancements – rather it shows broader levels of communications for even minor releases and beta program.
Commentary: Key Themes & Implications
Release cadence & momentum
- Elasticsearch: Shifted into the 9.x major version in April, then two minors by July & October. That shows a fairly aggressive pace.
- OpenSearch: Retained its 2.19.x line early in the year and launched 3.x in May, progressing quickly to 3.3.x by October 2025.
- Solr: Continued in the 9.x branch with meaningful feature releases (9.8 → 9.9) rather than a full major bump.
Vector search / GenAI readiness
- Elasticsearch 9.0 touts “binary quantization” improvements and general performance across vector workloads.
- OpenSearch 3.x expands vector + ML/observability features beyond the 2.19 line, aligning the platform with AI-driven workloads.
- Solr 9.9 added indexing-time vector embedding via external LLM services – a strong signal that Solr is embracing the AI-augmented search era.
Ops, scale & enterprise features
- Solr’s Cross-DC replication graduating from sandbox means that Solr is reinforcing its enterprise/HA credentials.
- Elasticsearch’s release of 9.0 introduced breaking changes (for example Java client requires Java 17) and ecosystem support (security plugins).
- OpenSearch continues its regular patch/minor cadence, with 3.0 released in May and 3.3.x arriving by October – signaling stability for long-term maintenance.
Buyer/Planner Guide: What to Do Now
- If you’re on Solr older than 9.8, you should evaluate upgrade urgency: 9.8 introduced Cross-DC replication officially; 9.9 adds vector/LLM embedding support.
- If you’re on Elasticsearch 8.x, now is the time to plan 9.x upgrade: factor in compatibility (Java client, breaking changes), test vector workloads, evaluate performance gains.
- If you’re on OpenSearch 2.17 or earlier, review 2.19.x features but plan for migration to the 3.x line (released in May and now at 3.3.x), which brings major performance, vector, and ML/observability improvements.
- Across all platforms: Vector/semantic/LLM-augmented search is now in production view. Your architecture needs to evaluate not just “full-text” queries but embedding workflows, quantization, hybrid search, and observability/insights.
- Ops/maintenance: Check your upgrade policies, support windows (e.g., Elasticsearch 8.19 vs 9.x), plugin ecosystem viability (security, observability, vector libraries) and multi-region / replication readiness.
Closing Thoughts
2025’s key takeaway: Search platforms aren’t just about inverted index and keyword – vector and generative workflows are now very real. The three major open‐source stacks are each evolving differently:
- Solr: steady, mature, enterprise-centric with vector/LLM capabilities entering.
- Elasticsearch: big version jump (9.x) and fast minor cadence — high feature velocity.
- OpenSearch: community-driven, strong vector/ML emphasis, and now firmly on the 3.x line (3.0 → 3.3.x).
For any organization building search or discovery experiences (internal or external) this year, the upgrade path, feature set and operational model of your platform choice will matter deeply. Use this timeline as a reference to map where you sit and where you’re headed.
To learn how Pureinsights helps organizations build and modernize AI-powered search applications, explore our Discovery Platform or request a free AI Search demo.
Additional Resources
- Top 7 Elasticsearch Pitfalls (and How to Avoid Them)
- Search Engine Rankings 2025
- Elasticsearch vs OpenSearch in 2025: What the Fork?
- Comparing Vector Search Solutions 2024
- Elasticsearch Consulting Services – Pureinsights
- Solr Consulting Services – Pureinsights
- OpenSearch Consulting Services – Pureinsights