G9v3-39A5B
G9v3-39A5B by AI9Stars is a 39B MoE model with 5B activated per token, dual Think and No Think modes, 131K context, and local self-hosted deployment.
Model Overview
Capabilities, design details, and architectural traits
G9v3-39A5B - Sparse MoE with dual Think modes
G9v3-39A5B is a Mixture-of-Experts causal language model from AI9Stars that activates only 5B of its 39B total parameters per token. The sparse design keeps inference cost close to a small dense model while retaining the capacity of a much larger one, making it practical for local and self-hosted deployment.
| Trait | Detail |
|---|---|
| Sparse MoE architecture | 39B total parameters with 5B activated per token |
| Dual Think modes | Think and No Think modes served from the same checkpoint via enable_thinking flag |
| 131K context window | Supports 131,072-token context length |
| Targeted use cases | Everyday assistant use, coding, tool-use workflows, and reasoning tasks |
| Preview release | Under active development with expected performance and capability updates |
Think and No Think from one checkpoint
The model supports both Think and No Think modes through the same checkpoint, toggled via the enable_thinking parameter in the chat template. Think mode uses temperature=1.0 with top_p=0.95, while No Think mode uses temperature=0.7 with top_p=0.95. This lets users switch between reasoning-intensive and fast-response behavior without loading a separate model.
Benchmark Performance
Independent evaluations · Artificial Analysis
Accuracy & Capability Details
Compare Models Side-by-Side
Evaluate specifications, pricing, and independent benchmark indices
| Model Details | |||
|---|---|---|---|
| General Info | |||
| Provider | AI9Stars | Anthropic | Anthropic |
| Release Date | August 3, 2026 | September 22, 2026 | September 28, 2026 |
| Knowledge Cutoff | - | - | Jun 2026 |
| Context & Limits | |||
| Context Window | 131K | 1M Best Context Window | 1M Best Context Window |
| Pricing (per 1M tokens) | |||
| Input Pricing | Free Best Input Pricing | $4 | $2 |
| Output Pricing | Free Best Output Pricing | $20 | $10 |
| Modalities | |||
| Inputs | text | textimagefile | textimagefile |
| Outputs | text | text | text |
| Benchmarks (0-100) | |||
| Intelligence Index | 31.6 | 57.6 Best Intelligence Index | 56.0 |
| Coding Index | 31.7 | - | - |
| Agentic Index | 28.7 | - | - |
Humanity's Last Exam
Extremely difficult logical reasoning and knowledge.
Long Context Reasoning
Logical reasoning over long context windows.
SciCode Benchmark
Scientific coding and mathematical modeling.
Independent evaluation data provided by Artificial Analysis. To view the latest benchmarks and full details, visit their official site.
Explore more from AI9Stars
Other models by AI9Stars
Top AI Models
Leading alternatives by intelligence score
