Leaderboard / Microsoft (Phi)
FrogNano 4B
Microsoft (Phi) · FrogNano · released 2026-09-17 · data checked 2026-11-01
FrogNano 4B is a 4.7B-parameter dense model from Microsoft (Phi) (United States), released 2026-09-17 under the MIT license. It doesn't have enough reported benchmarks for an index score yet. It needs about 3.7 GB of memory at 4-bit with an 8K-token context.
Small coding agent for repository-level software engineering (SWE-bench Pro: 37.6).
Benchmarks
Each bar shows this model's score. The tick marks the best tiny model on that benchmark. Sources: 1 reported by Microsoft (Phi). How we label sources.
Specifications
- Lab
- Microsoft (Phi) (US · United States)
- Based on
- Qwen3.5 4B (fine-tune or distillation)
- Architecture
- Dense transformer
- Total parameters
- 4.7B
- Active parameters
- All (dense)
- Size class
- Tiny
- Context window
- 262K tokens
- Input modalities
- text
- Reasoning
- Reasoning (thinking)
- License
- MIT
- Openness
- Open weights
- Memory to run
- ~3.7 GB at 4-bit, 8K context (more sizes)
- Hugging Face
- 2.7K downloads (30 days) · 178 likes
Compare with
- FrogNano 4B vs Gemma 3 4B (4.3B, index 17.5)
- FrogNano 4B vs Gemma 4 E2B (5.1B, index 31.9)
- FrogNano 4B vs Nanbeige4.2 3B (4.2B, index 70.4)
Sources
Something wrong here? Report it.
Run it locally
Uses bartowski/FrogNano-4B-2609-GGUF (Q4_K_M). Swap the tag for another size: BF16, IQ2_M, IQ3_M, IQ3_XS, IQ3_XXS, IQ4_NL, IQ4_XS, Q2_K….
ollama run hf.co/bartowski/FrogNano-4B-2609-GGUF:Q4_K_Mllama-server -hf bartowski/FrogNano-4B-2609-GGUF:Q4_K_Mvllm serve microsoft/FrogNano-4B-2609Memory needed: weights, context cache and runtime buffers.
| Quantization | 4K context | 32K context | 256K context |
|---|---|---|---|
| 16-bit F16 | 10 GB | 11 GB | 18 GB |
| 8-bit Q8_0 | 5.7 GB | 6.6 GB | 14 GB |
| 4-bit Q4_K_M | 3.6 GB | 4.5 GB | 12 GB |