whatmodel.sh / Jamba 1.5 Mini
Jamba 1.5 Mini
↗ share
+ contribute
Jamba 1.5 Mini is an open-weights language model from AI21 Labs, released August 2024. Its published size is 52B parameters (12B active per token). Licence: jamba-open-model-license.
Catalogue identity notes
Comparison-only. Arena entry 'jamba-1.5-mini' (text). LMArena's leaderboard table (arena-leaderboard space @6f2eeeb) labels 'jamba-1.5-mini' as 'Jamba-1.5-Mini' (AI21 Labs, Jamba Open), linking https://www.ai21.com/jamba. Maker repo ai21labs/AI21-Jamba-Mini-1.5 @5a844bb: licence tag other (jamba-open-model-license). Parameters: AI21 docs: Jamba Mini '52B parameters (12B active)' (HF safetensors metadata 51.6B). Released 2024-08: month the maker's Hugging Face repo was created (createdAt 2024-08-19). AI21 API snapshot jamba-mini-1.5-2024-08, deprecated 2025-05-06.
$ how it compares
The LLMs ranked either side of Jamba 1.5 Mini, out of 828 with a score, then more from AI21 Labs.
| # | model | score | where | ||
|---|---|---|---|---|---|
| 489 |  GPT-3.5 Turbo (0125) | 46±1 | api $0.75/M | ||
| 490 |  Mistral Large (2402) | 46±1 | comparison only | ||
| 491 | Jamba 1.5 Mini | 46±1 | local | ||
| 492 |  MedGemma 4B IT | 46±3 | local | ||
| 493 |  OLMo 2 13B Instruct (1124) | 46±1 | local | ||
| more from AI21 Labs | |||||
| 406 | Jamba 1.5 Large | 50±1 | local | ||
$ what it's good at
agent tasks▁▁▃▆█▆▅▅▅▄▃▁ #141 of 300 * · 3.0% of its score · 1 test
Beats Qwen3 235B-A22B Thinking 2507, o4-mini, o1 (2024-12-17) here, though they rank higher overall. Trails SmolLM3 3B here, though they rank lower overall.
| test | result | vs the others in its run | counts |
|---|---|---|---|
| BFCL (version not stated) lab | 80.7% ✎ | ▄▆▄▂·█▆·▄··▂ beats 8/17 | 3.0% |
coding█··········█ #2 of 2 * · 11% of its score · 1 test
chat & writing▁▁▁▃▅▄▅█▆▆▇▃ #281 of 393 · 23% of its score · 3 tests
Beats Ministral 3 14B Reasoning 2512, Gemma 4 E4B, Ministral 3 8B here, though they rank higher overall. Trails Flan-T5 XXL, LFM2 700M, EXAONE 3.5 2.4B Instruct here, though they rank lower overall.
| test | result | vs the others in its run | counts |
|---|---|---|---|
| LMArena (text) | 1239.7 ✎ | ▁▁▁▂▃▃▃▆▆▇█▃ beats 89/379 | 14% |
| Arena-Hard (version/judge unspecified) lab | 46.1% ✎ | ▃▃▂▅▃▆▅█▅▅▂▃ beats 7/30 | 5.2% |
| IFEval (aggregation unspecified) lab | 75.8% ✎ | ▁▁▃▁▂▄▄▆█▃▂▁ beats 46/141 | 3.8% |
math▁▂▃▅▆▅▄█▄▁▁▁ #311 of 398 * · 13% of its score · 1 test
Beats Athene-70B, OLMo 2 32B Instruct (0325), Hermes 2 Theta Llama-3 70B here, though they rank higher overall. Trails Gemma 7B (base), DeepSeek R1 Distill Qwen 1.5B, Tulu 2 DPO 13B here, though they rank lower overall.
other tests8 not in a task · 49% of its score
| test | result | vs the others in its run | counts |
|---|---|---|---|
| BIG-Bench Hard lab | 53.4% ✎ | ▁▁▁▁▂▄▅▅█▃▄▁ beats 56/129 | 13% |
| GPQA (subset unspecified) lab | 32.3% ✎ | ▂▁▂▂▄▃▆▃▅█▄▃ beats 28/104 | 10% |
| MMLU lab | 69.7% ✎ | ▁▁▁▁▁▂▃▅█▃▆▃ beats 111/183 | 10% |
| MMLU-Pro lab | 39.8% ✎ | ▁▁·▁▁▁▃▂▄▅█▁ beats 25/188 | 7.3% |
| RULER 64k | 92.8% ✎ | ▁▁·▁·▂▃▄█▂·▃ beats 22/22 | 3.6% |
| ARC Challenge lab | 85.7% ✎ | ▆▁▁▂▂▅▇█▄▃▃▂ beats 93/97 | 2.7% |
| RULER 128k | 90.0% ✎ | ▄··▂▄▂▆██▄▄▃ beats 21/21 | 1.3% |
| Vectara hallucination leaderboard, 2025 dataset (HHEM-2.3, factual consistency rate) | 97.1% ✎ | ▁▁▂▃▄▄█▆▃▂▂▁ beats 106/140 | 1.0% |
$ where to run it
llama.cpp · recommended
- Install llama.cpp · Homebrew and winget installs name the commands llama-cli and llama-server instead.macOS / Linux
curl -LsSf https://llama.app/install.sh | shWindowsirm https://llama.app/install.ps1 | iex - Run Jamba 1.5 Mini · needs ~31.2 GB · from huggingface.co/mradermacherchat in the terminal
llama cli -hf mradermacher/AI21-Jamba-Mini-1.5-GGUF:Q4_K_M - Or start a local server and open the address it prints (http://localhost:9931)
llama serve -hf mradermacher/AI21-Jamba-Mini-1.5-GGUF:Q4_K_M
LM Studio
- Install LM Studio · Open the app once before using the lms command.
- Run Jamba 1.5 Mini · needs ~31.2 GB · from huggingface.co/mradermacher
- Or download it in the terminal
lms get https://huggingface.co/mradermacher/AI21-Jamba-Mini-1.5-GGUF@Q4_K_M -y - Then run lms ls to see its name, and lms chat followed by that name
Ollama
- Install Ollama
download it (macOS / Windows) ↗
Linuxcurl -fsSL https://ollama.com/install.sh | sh - Run Jamba 1.5 Mini · needs ~31.2 GB · from huggingface.co/mradermacherchat in the terminal
ollama run hf.co/mradermacher/AI21-Jamba-Mini-1.5-GGUF:Q4_K_M
Other ways to run it
- Run it as published · needs ~103.14 GB, NVIDIA only · from github.com/vllm-project
vllm serve ai21labs/AI21-Jamba-Mini-1.5
Official weights and documentation: huggingface.co/ai21labs ↗
$ how the score was worked out · Jamba-1.5: Hybrid Transformer… 79%, LMArena leaderboard dataset 14%, NVIDIA RULER repository READM… 5% and 1 more
Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works
| source | results | share of the score |
|---|---|---|
 Jamba-1.5: Hybrid Transformer-Mamba Models at Scale (AI21, arXiv 2408.12570v1) · lab-reported | 10 | 79% |
 LMArena leaderboard dataset | 1 | 14% |
 NVIDIA RULER repository README results table | 2 | 5% |
| Vectara hallucination leaderboard (vectara/hallucination-leaderboard repository) | 1 | 1% |
| prior · pull toward the middle | 1% | |
| weighted average | = 46.4 |
$ improve this page
Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.
data updated · sources and licences · json