- Size
- 16.0B parameterscounted from the safetensors header on the Hub
- Licence
- unknown, BigCode OpenRAIL-MOur reading, check the licence: A responsible AI licence: commercial use is allowed, but the licence lists uses that are not allowed, and you must pass those use restrictions on to anyone you give the model or a changed version to.
- On the Hub since
- 20 Feb 2024the date the repo was created
- Context length
- 16k tokensfrom the model's config
- Kind
- Base model (no base model declared on the Hub)Starcoder2ForCausalLM
- On Hugging Face
- bigcode/starcoder2-15b14k downloads in the last 30 days, 684 likes
Scores
- default setting47.2%Qwen2.5-Coder Technical Report
Epoch's name: starcoder2-15b. Best closed model on this set: gpt-3.5-turbo-1106, 87.4% (OpenAI). All 134 results
- Shots: 457.7%Qwen2.5-Coder Technical Report
Epoch's name: starcoder2-15b. Best closed model on this set: gpt-4-0314, 92.0% (OpenAI). All 170 results
- default setting64.1%Qwen2.5-Coder Technical Report
Epoch's name: starcoder2-15b. Best closed model on this set: gpt-4o-2024-11-20, 88.1% (OpenAI). All 217 results
- default setting64.3%Qwen2.5-Coder Technical Report
Epoch's name: starcoder2-15b. Best closed model on this set: claude-3-opus-20240229, 88.5% (Anthropic). All 137 results
Claimed by the maker Claimed
- CruxEval-I pass@148.1
- DS-1000 pass@133.8
- GSM8K (PAL) accuracy65.1
- HumanEval+ pass@137.8
- HumanEval pass@146.3
- RepoBench-v1.1 edit-smiliarity74.08
Values as the card gives them, unscaled. Source: the model card on the Hugging Face Hub. Claimed scores are never mixed with independent ones.
Other views, which we link to and do not copy: look this model up on Artificial Analysis and the Arena leaderboard.
Independent scores from Epoch AI (CC-BY 4.0), fetched 19 Nov 2026: Epoch AI, ‘Capabilities & benchmarking’. Published online at epoch.ai. Retrieved from ‘https://epoch.ai/benchmarks’ [online resource]. Each row links its own source. What the badges mean.
Run it yourself
Memory needed = weights + context cache + 0.50 GB for the runtime. Context cache at 8k tokens: 0.63 GB, worked out from the model's config (bigcode/starcoder2-15b), at 16-bit precision. Choose a machine above to see a verdict for each format. Method
GGUF for llama.cpp, Ollama, LM Studio and others
second-state/StarCoder2-15B-GGUF (converted by second-state), 2.9k downloads. Sizes are the real files on the Hub, all parts added up.
| Format | Size | Memory needed | Get it |
|---|---|---|---|
| Q2_K | 5.8 GBfile size | 6.9 GB5.8 + 0.63 + 0.50 | Commandsllama.cpp llama-server -hf second-state/StarCoder2-15B-GGUF:Q2_KOllama ollama run hf.co/second-state/StarCoder2-15B-GGUF:Q2_KDownload hf download second-state/StarCoder2-15B-GGUF --include "starcoder2-15b-Q2_K.gguf" --local-dir StarCoder2-15B-GGUF |
| Q3_K_S | 6.5 GBfile size | 7.6 GB6.5 + 0.63 + 0.50 | Commandsllama.cpp llama-server -hf second-state/StarCoder2-15B-GGUF:Q3_K_SOllama ollama run hf.co/second-state/StarCoder2-15B-GGUF:Q3_K_SDownload hf download second-state/StarCoder2-15B-GGUF --include "starcoder2-15b-Q3_K_S.gguf" --local-dir StarCoder2-15B-GGUF |
| Q3_K_M | 7.5 GBfile size | 8.7 GB7.5 + 0.63 + 0.50 | Commandsllama.cpp llama-server -hf second-state/StarCoder2-15B-GGUF:Q3_K_MOllama ollama run hf.co/second-state/StarCoder2-15B-GGUF:Q3_K_MDownload hf download second-state/StarCoder2-15B-GGUF --include "starcoder2-15b-Q3_K_M.gguf" --local-dir StarCoder2-15B-GGUF |
| Q3_K_L | 8.3 GBfile size | 9.5 GB8.3 + 0.63 + 0.50 | Commandsllama.cpp llama-server -hf second-state/StarCoder2-15B-GGUF:Q3_K_LOllama ollama run hf.co/second-state/StarCoder2-15B-GGUF:Q3_K_LDownload hf download second-state/StarCoder2-15B-GGUF --include "starcoder2-15b-Q3_K_L.gguf" --local-dir StarCoder2-15B-GGUF |
| Q4_0 | 8.4 GBfile size | 9.6 GB8.4 + 0.63 + 0.50 | Commandsllama.cpp llama-server -hf second-state/StarCoder2-15B-GGUF:Q4_0Ollama ollama run hf.co/second-state/StarCoder2-15B-GGUF:Q4_0Download hf download second-state/StarCoder2-15B-GGUF --include "starcoder2-15b-Q4_0.gguf" --local-dir StarCoder2-15B-GGUF |
| Q4_K_S | 8.6 GBfile size | 9.7 GB8.6 + 0.63 + 0.50 | Commandsllama.cpp llama-server -hf second-state/StarCoder2-15B-GGUF:Q4_K_SOllama ollama run hf.co/second-state/StarCoder2-15B-GGUF:Q4_K_SDownload hf download second-state/StarCoder2-15B-GGUF --include "starcoder2-15b-Q4_K_S.gguf" --local-dir StarCoder2-15B-GGUF |
| Q4_K_M | 9.2 GBfile size | 10.3 GB9.2 + 0.63 + 0.50 | Commandsllama.cpp llama-server -hf second-state/StarCoder2-15B-GGUF:Q4_K_MOllama ollama run hf.co/second-state/StarCoder2-15B-GGUF:Q4_K_MDownload hf download second-state/StarCoder2-15B-GGUF --include "starcoder2-15b-Q4_K_M.gguf" --local-dir StarCoder2-15B-GGUF |
| Q5_0 | 10.3 GBfile size | 11.4 GB10.3 + 0.63 + 0.50 | Commandsllama.cpp llama-server -hf second-state/StarCoder2-15B-GGUF:Q5_0Ollama ollama run hf.co/second-state/StarCoder2-15B-GGUF:Q5_0Download hf download second-state/StarCoder2-15B-GGUF --include "starcoder2-15b-Q5_0.gguf" --local-dir StarCoder2-15B-GGUF |
| Q5_K_S | 10.3 GBfile size | 11.4 GB10.3 + 0.63 + 0.50 | Commandsllama.cpp llama-server -hf second-state/StarCoder2-15B-GGUF:Q5_K_SOllama ollama run hf.co/second-state/StarCoder2-15B-GGUF:Q5_K_SDownload hf download second-state/StarCoder2-15B-GGUF --include "starcoder2-15b-Q5_K_S.gguf" --local-dir StarCoder2-15B-GGUF |
| Q5_K_M | 10.6 GBfile size | 11.8 GB10.6 + 0.63 + 0.50 | Commandsllama.cpp llama-server -hf second-state/StarCoder2-15B-GGUF:Q5_K_MOllama ollama run hf.co/second-state/StarCoder2-15B-GGUF:Q5_K_MDownload hf download second-state/StarCoder2-15B-GGUF --include "starcoder2-15b-Q5_K_M.gguf" --local-dir StarCoder2-15B-GGUF |
| Q6_K | 12.2 GBfile size | 13.3 GB12.2 + 0.63 + 0.50 | Commandsllama.cpp llama-server -hf second-state/StarCoder2-15B-GGUF:Q6_KOllama ollama run hf.co/second-state/StarCoder2-15B-GGUF:Q6_KDownload hf download second-state/StarCoder2-15B-GGUF --include "starcoder2-15b-Q6_K.gguf" --local-dir StarCoder2-15B-GGUF |
| Q8_0 | 15.8 GBfile size | 16.9 GB15.8 + 0.63 + 0.50 | Commandsllama.cpp llama-server -hf second-state/StarCoder2-15B-GGUF:Q8_0Ollama ollama run hf.co/second-state/StarCoder2-15B-GGUF:Q8_0Download hf download second-state/StarCoder2-15B-GGUF --include "starcoder2-15b-Q8_0.gguf" --local-dir StarCoder2-15B-GGUF |
MLX for Apple silicon
| Format | Size | Memory needed | Repo |
|---|---|---|---|
| MLX 4bitmlx-community/bigcode-starcoder2-15b-4bit | 8.4 GBfile size | 9.5 GB8.4 + 0.63 + 0.50 | Download page |
| MLX 8bitmlx-community/bigcode-starcoder2-15b-8bit | 15.8 GBfile size | 16.9 GB15.8 + 0.63 + 0.50 | Download page |
Original weights size from the safetensors header, for vLLM, SGLang, Transformers
| Format | Size | Memory needed | Repo |
|---|---|---|---|
| F32original weights | 59.4 GBfile size | 60.6 GB59.4 + 0.63 + 0.50 |
hf download bigcode/starcoder2-15b --local-dir starcoder2-15bOther conversions (4; sizes not read)
- RedHatAI/starcoder2-15b-quantized.w8a16 Other, 124 downloads
- lainlives/starcoder2-15b-bnb-4bit bitsandbytes, 4bit, 62 downloads
- shavera/starcoder2-15b-w4-autoawq-gemm AWQ, 28 downloads
- onekq-ai/starcoder2-15b-bnb-4bit bitsandbytes, 4bit, 0 downloads
The Hub tags 22 quantised repos as conversions of this model; we read the most downloaded and show the maker's own first.
Run it by API
No provider on Hugging Face Inference Providers serves this model (as of 19 Nov 2026, 13:58 UTC). Other hosts may; we only list what the router reports.
Not built yet: one key for every open model, with BlockRun, and our own tested-quality score per provider. Until then, precision and measured speed are the best guide to a provider here.
Fine-tunes of this model
Repos that declare this model as their base, most downloaded first. Listing is not a recommendation; unreviewed ones have not been checked.
- bigcode/starcoder2-15b-instruct-v0.116.0B1.8kUnreviewed
Planned for this page Not built yet
- Try it in private chat
- A link to Solana AI's confidential chat for the models it serves. This model is not listed as served there.
- Fine-tune it
- Fine-tuning a few base models from this page and publishing the result here.
- Creators and earnings
- Pages for people who publish fine-tunes and adapters of this model, and what they earn.
- Fund and bounties
- Bounties for new open models, paid from a public fund. How the fund will work.
- Contribute compute
- How to give compute to open model work, with Akash.
Model data from the Hugging Face Hub 19 Nov 2026, 13:43 UTC. The repo. Found a mistake on this page? Tell us.