FiveForge v2.0 is our biggest update since launch. We rebuilt the AI infrastructure, added community features, and shipped the most-requested features from our users.
Here's everything that's new and why it matters.
New Model Lineup
v2.0 introduces a completely reworked model architecture with two providers:
FiveForge AI Provider
| Model | Best For | Availability |
|---|---|---|
| Ultra | Complex multi-file resource generation | GO |
| Flash | Quick syntax and API questions | Free + GO |
| Reason | Debugging and root cause analysis | GO |
| Code | Lua, JS, SQL generation | GO |
| Turbo | Balanced speed and quality | GO |
Open Models Provider
| Model | Best For | Availability |
|---|---|---|
| Prime | General-purpose development help | GO |
| Nova | Lightweight quick answers | GO |
The key change: Flash is now free. This makes FiveForge accessible to every FiveM developer, even those not ready for a paid plan. Flash handles syntax questions, quick code reviews, and simple script generation — enough to be useful day-to-day.
When to Use Each Model
Getting the right model for the task makes a real difference. Here's the rule of thumb:
- Scaffolding a new resource? Use Code — it's optimized for generating production-ready code with proper structure
- Debugging a cryptic error? Use Reason — it steps through framework event chains to find root causes
- Everyday questions? Use Flash (free) — it's fast and handles most common tasks
- Complex multi-file changes? Use Ultra — it has the largest context window
Community Features
This was the most-requested category by a wide margin. Here's what we shipped:
Invite System with Analytics
Track invites per Discord invite code with detailed stats: total invites, valid invites, pending verifications, fake account detection, and members who left. The leaderboard ranks by net invites (valid minus left), so invite grinding isn't rewarded.
Automated Badge System
Badges sync automatically to Discord roles, giving your community visible progression:
| Badge | Requirement | Discord Role |
|---|---|---|
| Bronze Recruiter | 10 invites | @Bronze |
| Silver Recruiter | 25 invites | @Silver |
| Gold Recruiter | 50 invites | @Gold |
| Diamond Recruiter | 100 invites | @Diamond |
| Community Legend | 250 invites | @Legend |
| First Win | 1 giveaway win | @Lucky |
| Lucky Winner | 5 giveaway wins | @Lucky+ |
| GO Member | Active subscription | @GO |
Giveaways with Two Modes
- Random Draw — any community member can enter
- Invite Rank — weighted entries based on invite count
GO-exclusive giveaways offer better prizes, creating a tangible benefit for subscribers.
Hall of Fame
Top contributors and invite champions are featured on the community site, with automatic updates as the leaderboard changes.
Infrastructure Upgrades Under the Hood
These changes don't have a UI, but they make the biggest difference in day-to-day use:
- Automatic model failover — If a model hits its rate limit, FiveForge seamlessly falls back to an available model. No more "service unavailable" errors
- Rolling usage window — Your daily limits reset 24 hours after your last request, not at midnight. This means if you work late, you don't lose half your daily quota
- Improved sidecar stability — The local process that handles AI communication is more resilient to crashes and memory leaks. It now auto-restarts if it detects instability
- Faster response times — Optimized routing reduces latency across all models by roughly 35%
What v2.0 Enables
v2.0 isn't just a feature release — it's a foundation for what comes next:
- Multi-file editing — edit multiple resource files simultaneously with AI awareness of cross-file dependencies
- Live debugging — AI that connects to your server console in real-time
- Architecture review — full server resource audit with prioritized recommendations
- Team collaboration — shared sessions and templates for development teams
What Users Are Saying
"The Flash model being free changed how I work. I used to only ask FiveForge for big tasks. Now I use it for every little thing — syntax checks, quick code reviews, even writing documentation." — FiveForge user, 2 weeks on v2.0
"The model failover saved me during a late-night dev session. One model hit its limit and I didn't even notice the switch." — FiveForge user, v2.0 beta tester
Ready to try v2.0? Download the latest version or upgrade to GO for the full experience.
ZiadLead Developer
Ziad is the lead developer of FiveForge. He builds AI systems that read codebases and get work done.
Stay updated
Get the latest AI development guides and FiveForge updates delivered to your inbox.
Build faster with FiveForge
An AI agent for complex work. Plan, build, debug, and automate projects — all from your desktop.