Simon Willison reviews Ornith-1.0, a new series of open-weights models from DeepReinforce designed for agentic coding. Built on Gemma 4 and Qwen 3.5, these models range from 9B to 397B parameters and demonstrate high proficiency in coding benchmarks and tool-use tasks.
- Available in 9B Dense, 31B Dense, 35B MoE, and 397B MoE variants
- Capable of navigating codebases and executing complex tool calls
- Built on Apache 2.0 licensed base models
- Available in 9B Dense, 31B Dense, 35B MoE, and 397B MoE variants
- Capable of navigating codebases and executing complex tool calls
- Built on Apache 2.0 licensed base models
An exploration of high-performing small language models with under 7 billion parameters that can run locally on consumer hardware like laptops and smartphones. The article explains how advancements in training data quality, model distillation from larger frontier models, and architectural improvements like Mixture-of-Experts have enabled these compact models to compete with much larger versions on reasoning benchmarks. It provides a curated guide of top available models on Hugging Face, detailing their specific strengths, benchmark performance, and providing Python code for implementation.
Key models covered:
- Qwen3.5-4B for multilingual tasks and long context windows
- Microsoft Phi-4-mini-instruct for reasoning-heavy English workloads
- Google Gemma 3 4B IT for coding and mathematics
- Google Gemma 3n E4B for efficient mobile and on-device deployment
- Meta Llama 3.2 3B Instruct for tool calling and community support
- SmolLM3-3B for research transparency and open-source projects
- DeepSeek-R1-Distill-Qwen-1.5B for lightweight reasoning on edge devices
- Qwen3-0.6B for ultra-constrained hardware and text classification
Key models covered:
- Qwen3.5-4B for multilingual tasks and long context windows
- Microsoft Phi-4-mini-instruct for reasoning-heavy English workloads
- Google Gemma 3 4B IT for coding and mathematics
- Google Gemma 3n E4B for efficient mobile and on-device deployment
- Meta Llama 3.2 3B Instruct for tool calling and community support
- SmolLM3-3B for research transparency and open-source projects
- DeepSeek-R1-Distill-Qwen-1.5B for lightweight reasoning on edge devices
- Qwen3-0.6B for ultra-constrained hardware and text classification
This discussion details performance benchmarks of llama.cpp on an NVIDIA DGX Spark, including tests for various models (gpt-oss-20b, gpt-oss-120b, Qwen3, Qwen2.5, Gemma, GLM) with different context depths and batch sizes.
A detailed comparison of the architectures of recent large language models (LLMs) including DeepSeek-V3, OLMo 2, Gemma 3, Mistral Small 3.1, Llama 4, Qwen3, SmolLM3, and Kimi 2, focusing on key design choices and their impact on performance and efficiency.
1. **DeepSeek V3/R1**:
- Uses Multi-Head Latent Attention (MLA) and Mixture-of-Experts (MoE) for efficiency.
- MLA compresses key and value tensors to reduce KV cache memory usage.
- MoE activates only a subset of experts per token, improving inference efficiency.
2. **OLMo 2**:
- Focuses on transparency in training data and code.
- Uses RMSNorm layers placed after attention and feed-forward modules (Post-Norm).
- Introduces QK-Norm, an additional RMSNorm layer applied to queries and keys inside the attention mechanism.
3. **Gemma 3**:
- Employs sliding window attention to reduce memory requirements in the KV cache.
- Uses a 5:1 ratio of sliding window attention to global attention layers.
- Combines Pre-Norm and Post-Norm RMSNorm layers around the attention module.
4. **Mistral Small 3.1**:
- Outperforms Gemma 3 27B on several benchmarks while being faster.
- Uses a standard architecture with a custom tokenizer and reduced KV cache and layer count.
5. **Llama 4**:
- Adopts an MoE approach similar to DeepSeek V3 but with fewer, larger experts.
- Alternates MoE and dense modules in every other transformer block.
6. **Qwen3**:
- Comes in both dense and MoE variants.
- Dense models are easier to fine-tune and deploy, while MoE models are optimized for scaling inference.
7. **SmolLM3**:
- Uses No Positional Embeddings (NoPE), omitting explicit positional information injection.
- NoPE improves length generalization, meaning performance deteriorates less with increased sequence length.
8. **Kimi K2 and Kimi K2 Thinking**:
- Uses a variant of the Muon optimizer over AdamW.
- Kimi K2 Thinking extends the context size to 256k tokens.
9. **GPT-OSS**:
- OpenAI's first open-weight models since GPT-2.
- Uses sliding window attention and a width-versus-depth trade-off.
10. **Grok 2.5**:
- Uses a small number of large experts and a shared expert module.
- Reflects an older trend in MoE architectures.
11. **GLM-4.5**:
- Comes in two variants: a 355-billion-parameter model and a more compact 106-billion-parameter version.
- Uses a shared expert and starts with several dense layers before introducing MoE blocks.
12. **Qwen3-Next**:
- Introduces a Gated DeltaNet + Gated Attention hybrid mechanism.
- Uses Multi-Token Prediction (MTP) for efficiency.
13. **MiniMax-M2**:
- Uses per-layer QK-Norm and partial RoPE.
- More "sparse" than Qwen3, with fewer active experts per token.
14. **Kimi Linear**:
- Modifies the linear attention mechanism with Kimi Delta Attention (KDA).
- Combines Gated DeltaNet with Multi-Head Latent Attention (MLA).
15. **Olmo 3 Thinking**:
- Uses sliding window attention and YaRN for context extension.
- Comes in base, instruct, and reasoning variants.
16. **DeepSeek V3.2**:
- Adds a sparse attention mechanism to improve efficiency.
- On par with GPT-5.1 and Gemini 3.0 Pro on certain benchmarks.
17. **Mistral 3**:
- First MoE model since Mixtral in 2023.
- Partnered with NVIDIA for optimization on Blackwell chips.
18. **Nemotron 3**:
- A Transformer-Mamba hybrid architecture.
- Interleaves Mamba-2 sequence-modeling blocks with sparse MoE feed-forward layers.
19. **Xiaomi MiMo-V2-Flash**:
- Uses sliding window attention in a 5:1 ratio with global attention.
- Employs multi-token prediction (MTP) for efficiency.
20. **Arcee AI Trinity Large**:
- Uses alternating local:global attention layers, NoPE, and gated attention.
- Introduces depth-scaled sandwich norm for training stability.
1. **DeepSeek V3/R1**:
- Uses Multi-Head Latent Attention (MLA) and Mixture-of-Experts (MoE) for efficiency.
- MLA compresses key and value tensors to reduce KV cache memory usage.
- MoE activates only a subset of experts per token, improving inference efficiency.
2. **OLMo 2**:
- Focuses on transparency in training data and code.
- Uses RMSNorm layers placed after attention and feed-forward modules (Post-Norm).
- Introduces QK-Norm, an additional RMSNorm layer applied to queries and keys inside the attention mechanism.
3. **Gemma 3**:
- Employs sliding window attention to reduce memory requirements in the KV cache.
- Uses a 5:1 ratio of sliding window attention to global attention layers.
- Combines Pre-Norm and Post-Norm RMSNorm layers around the attention module.
4. **Mistral Small 3.1**:
- Outperforms Gemma 3 27B on several benchmarks while being faster.
- Uses a standard architecture with a custom tokenizer and reduced KV cache and layer count.
5. **Llama 4**:
- Adopts an MoE approach similar to DeepSeek V3 but with fewer, larger experts.
- Alternates MoE and dense modules in every other transformer block.
6. **Qwen3**:
- Comes in both dense and MoE variants.
- Dense models are easier to fine-tune and deploy, while MoE models are optimized for scaling inference.
7. **SmolLM3**:
- Uses No Positional Embeddings (NoPE), omitting explicit positional information injection.
- NoPE improves length generalization, meaning performance deteriorates less with increased sequence length.
8. **Kimi K2 and Kimi K2 Thinking**:
- Uses a variant of the Muon optimizer over AdamW.
- Kimi K2 Thinking extends the context size to 256k tokens.
9. **GPT-OSS**:
- OpenAI's first open-weight models since GPT-2.
- Uses sliding window attention and a width-versus-depth trade-off.
10. **Grok 2.5**:
- Uses a small number of large experts and a shared expert module.
- Reflects an older trend in MoE architectures.
11. **GLM-4.5**:
- Comes in two variants: a 355-billion-parameter model and a more compact 106-billion-parameter version.
- Uses a shared expert and starts with several dense layers before introducing MoE blocks.
12. **Qwen3-Next**:
- Introduces a Gated DeltaNet + Gated Attention hybrid mechanism.
- Uses Multi-Token Prediction (MTP) for efficiency.
13. **MiniMax-M2**:
- Uses per-layer QK-Norm and partial RoPE.
- More "sparse" than Qwen3, with fewer active experts per token.
14. **Kimi Linear**:
- Modifies the linear attention mechanism with Kimi Delta Attention (KDA).
- Combines Gated DeltaNet with Multi-Head Latent Attention (MLA).
15. **Olmo 3 Thinking**:
- Uses sliding window attention and YaRN for context extension.
- Comes in base, instruct, and reasoning variants.
16. **DeepSeek V3.2**:
- Adds a sparse attention mechanism to improve efficiency.
- On par with GPT-5.1 and Gemini 3.0 Pro on certain benchmarks.
17. **Mistral 3**:
- First MoE model since Mixtral in 2023.
- Partnered with NVIDIA for optimization on Blackwell chips.
18. **Nemotron 3**:
- A Transformer-Mamba hybrid architecture.
- Interleaves Mamba-2 sequence-modeling blocks with sparse MoE feed-forward layers.
19. **Xiaomi MiMo-V2-Flash**:
- Uses sliding window attention in a 5:1 ratio with global attention.
- Employs multi-token prediction (MTP) for efficiency.
20. **Arcee AI Trinity Large**:
- Uses alternating local:global attention layers, NoPE, and gated attention.
- Introduces depth-scaled sandwich norm for training stability.
A Reddit thread discussing preferred local Large Language Model (LLM) setups for tasks like summarizing text, coding, and general use. Users share their model choices (Gemma, Qwen, Phi, etc.) and frameworks (llama.cpp, Ollama, EXUI) along with potential issues and configurations.
| **Model** | **Use Cases** | **Size (Parameters)** | **Approx. VRAM (Q4 Quantization)** | **Approx. RAM (Q4)** | **Notes/Requirements** |
|----------------|---------------------------------------------------|------------------------|-----------------------------------|---------------------|-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| **Gemma 3 (Meta)** | Summarization, conversational tasks, image recognition, translation, simple writing | 3B, 4B, 7B, 8B, 12B, 27B+ | 2-4GB (3B), 4-6GB (7B), 8-12GB (12B) | 4-8GB (3B), 8-12GB (7B), 16-24GB (12B) | Excellent performance for its size. Recent versions have had memory leak issues (see Reddit post – use Ollama 0.6.6 or later, but even that may not be fully fixed). QAT versions are highly recommended. |
| **Qwen 2.5 (Alibaba)** | Summarization, coding, reasoning, decision-making, technical material processing | 3.5B, 7B, 72B | 2-3GB (3.5B), 4-6GB (7B), 26-30GB (72B) | 4-6GB (3.5B), 8-12GB (7B), 50-60GB (72B) | Qwen models are known for strong performance. Coder versions specifically tuned for code generation. |
| **Qwen3 (Alibaba - upcoming)**| General purpose, likely similar to Qwen 2.5 with improvements | 70B | Estimated 25-30GB (Q4) | 50-60GB | Expected to be a strong competitor. |
| **Llama 3 (Meta)**| General purpose, conversation, writing, coding, reasoning | 8B, 13B, 70B+ | 4-6GB (8B), 7-9GB (13B), 25-30GB (70B) | 8-12GB (8B), 14-18GB (13B), 50-60GB (70B) | Current state-of-the-art open-source model. Excellent balance of performance and size. |
| **YiXin (01.AI)** | Reasoning, brainstorming | 72B | ~26-30GB (Q4) | ~50-60GB | A powerful model focused on reasoning and understanding. Similar VRAM requirements to Qwen 72B. |
| **Phi-4 (Microsoft)** | General purpose, writing, coding | 14B | ~7-9GB (Q4) | 14-18GB | Smaller model, good for resource-constrained environments, but may not match larger models in complexity. |
| **Ling-Lite** | RAG (Retrieval-Augmented Generation), fast processing, text extraction | Variable | Varies with size | Varies with size | MoE (Mixture of Experts) model known for speed. Good for RAG applications where quick responses are important. |
**Key Considerations:**
* **Quantization:** The VRAM and RAM estimates above are based on 4-bit quantization (Q4). Lower quantization (e.g., Q2) will reduce memory usage further, but *may* impact quality. Higher quantization (e.g., Q8, FP16) will increase quality but require significantly more memory.
* **Frameworks:** Popular frameworks for running these models locally include:
* **llama.cpp:** Highly optimized for CPU and GPU, especially on Apple Silicon.
* **Ollama:** Simplified setup and management of LLMs. (Be aware of the Gemma 3 memory leak issue!)
* **Text Generation WebUI (oobabooga):** Web-based interface with many features and customization options.
* **Hardware:** A dedicated GPU with sufficient VRAM is highly recommended for decent performance. CPU-only inference is possible but can be slow. More RAM is generally better, even if the model fits in VRAM.
* **Context Length:** The "40k" context mentioned in the Reddit post refers to the maximum number of tokens (words or sub-words) the model can process at once. Longer context lengths require more memory.
| **Model** | **Use Cases** | **Size (Parameters)** | **Approx. VRAM (Q4 Quantization)** | **Approx. RAM (Q4)** | **Notes/Requirements** |
|----------------|---------------------------------------------------|------------------------|-----------------------------------|---------------------|-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| **Gemma 3 (Meta)** | Summarization, conversational tasks, image recognition, translation, simple writing | 3B, 4B, 7B, 8B, 12B, 27B+ | 2-4GB (3B), 4-6GB (7B), 8-12GB (12B) | 4-8GB (3B), 8-12GB (7B), 16-24GB (12B) | Excellent performance for its size. Recent versions have had memory leak issues (see Reddit post – use Ollama 0.6.6 or later, but even that may not be fully fixed). QAT versions are highly recommended. |
| **Qwen 2.5 (Alibaba)** | Summarization, coding, reasoning, decision-making, technical material processing | 3.5B, 7B, 72B | 2-3GB (3.5B), 4-6GB (7B), 26-30GB (72B) | 4-6GB (3.5B), 8-12GB (7B), 50-60GB (72B) | Qwen models are known for strong performance. Coder versions specifically tuned for code generation. |
| **Qwen3 (Alibaba - upcoming)**| General purpose, likely similar to Qwen 2.5 with improvements | 70B | Estimated 25-30GB (Q4) | 50-60GB | Expected to be a strong competitor. |
| **Llama 3 (Meta)**| General purpose, conversation, writing, coding, reasoning | 8B, 13B, 70B+ | 4-6GB (8B), 7-9GB (13B), 25-30GB (70B) | 8-12GB (8B), 14-18GB (13B), 50-60GB (70B) | Current state-of-the-art open-source model. Excellent balance of performance and size. |
| **YiXin (01.AI)** | Reasoning, brainstorming | 72B | ~26-30GB (Q4) | ~50-60GB | A powerful model focused on reasoning and understanding. Similar VRAM requirements to Qwen 72B. |
| **Phi-4 (Microsoft)** | General purpose, writing, coding | 14B | ~7-9GB (Q4) | 14-18GB | Smaller model, good for resource-constrained environments, but may not match larger models in complexity. |
| **Ling-Lite** | RAG (Retrieval-Augmented Generation), fast processing, text extraction | Variable | Varies with size | Varies with size | MoE (Mixture of Experts) model known for speed. Good for RAG applications where quick responses are important. |
**Key Considerations:**
* **Quantization:** The VRAM and RAM estimates above are based on 4-bit quantization (Q4). Lower quantization (e.g., Q2) will reduce memory usage further, but *may* impact quality. Higher quantization (e.g., Q8, FP16) will increase quality but require significantly more memory.
* **Frameworks:** Popular frameworks for running these models locally include:
* **llama.cpp:** Highly optimized for CPU and GPU, especially on Apple Silicon.
* **Ollama:** Simplified setup and management of LLMs. (Be aware of the Gemma 3 memory leak issue!)
* **Text Generation WebUI (oobabooga):** Web-based interface with many features and customization options.
* **Hardware:** A dedicated GPU with sufficient VRAM is highly recommended for decent performance. CPU-only inference is possible but can be slow. More RAM is generally better, even if the model fits in VRAM.
* **Context Length:** The "40k" context mentioned in the Reddit post refers to the maximum number of tokens (words or sub-words) the model can process at once. Longer context lengths require more memory.
