Bielik-11B-v3.0-Instruct
speakleash Apache 2.0 out 12 months ago
11.2B parameters, text in, text out. 1 provider serves it from $0.40 per million input tokens.
Verified 2026-10-25 Rank #633 in LLMs · score 73.3 method v1 Sources 3 Downloads 30d 3.5k · #565 of 619 tracked
Where to run it
Live hosted endpoints, cheapest first, USD per million tokens. Ties break on output price, then provider name.
| # | Cheapest live row | In /1M | Out /1M | Context | Quant | Seen |
|---|---|---|---|---|---|---|
| 1 | Public AI | $0.40 | $0.40 | — | — | Oct 25 |
Where it's used
Hugging Face does not say who downloads a model or with what. These are the proxies: API providers listing it, quantized builds, fine-tunes/adapters/merges, and Spaces that load it.
Served by 1 API provider; 20 quantized builds for local use.
- As a service1Public AI
- On own hardware20quantized builds
- Built on106 fine-tunes · 4 adapters · 0 merges
- In apps5Hugging Face Spaces
Top Spaces running it: agentGreg/czy-bielik-zna-temat ♥ 1
Will it fit?
Weights only, an estimate: leave room for the KV cache. Method
Fits a 16 GB card at INT4
About 6.7 GB at INT4, 13 GB at FP8, 27 GB at BF16: weights only, so leave room for the KV cache.
Estimate params × bytes per param × 1.2 · how we estimate
What people built on it
30 derivatives on the Hub. Which build to run, by runtime: the most downloaded in each class.
30 derivatives on the Hub, most of them quantized
From the Hugging Face model tree; quantized builds in teal, fine-tunes, merges and adapters in greys.
Reference
Everything else, folded: the full tables, every source.
What the card says
1 cited sentence
Summary generated, cited
This fine-tuned model supports multilingual conversational tasks in languages including Polish, English, and Bulgarian.[1]
Generated 2026-10-07 15:13 UTC by parasail-gemma-4-26b-a4b-it; each numeral points at the fact it rests on. Not a verdict.
Use-case evidence
3 use cases
What it is good at 3 use cases
Labels are evidence, not verdicts: measured a third-party leaderboard row for this model; author-reported a number the model author published in the card or eval files; structural a deterministic signal from the name, config or chat template, or a decision-model estimate. Each chip links to its source. All use cases.
Facts
17 fields
Listed in LLMs
Facts
- Bielik-11B-v3.0-Instruct has 11.2B parameters. (safetensors metadata, observed 2026-10-24)
- License: Apache 2.0 (friction: none). Permissive; patent grant. (license file, observed 2026-10-24)
- Created 2025-11-07, last modified 2026-07-01. (HF API, observed 2026-10-24)
- Access is gated on Hugging Face (auto). (HF API, observed 2026-10-24)
| Pipeline tag | text-generation |
|---|---|
| Library | transformers |
| Architectures | LlamaForCausalLM |
| Model type | llama |
| Params by dtype | BF16: 11,168,796,672 |
| Size band | L |
| MoE | no |
| Gated | auto |
| Papers | arXiv:2505.02410, arXiv:2505.02550, arXiv:2402.13228, arXiv:2402.03300, arXiv:2503.20783, arXiv:2409.19256, arXiv:2601.11579 |
| License | Apache 2.0 (apache-2.0) |
| Hub snapshot | 2026-10-25 |
Servability
vLLM · SGLang
Servability
vLLM: supported · LlamaForCausalLM · 7050ef881ff0835f6890a47db1859a24c65481d6SGLang: supported · LlamaForCausalLM · 773cb938fe9484a8f7e7e46fb1b1fe96412c59ed
vLLM registry (7050ef881ff0835f6890a47db1859a24c65481d6) · SGLang registry (773cb938fe9484a8f7e7e46fb1b1fe96412c59ed) · Matched by architecture class name against each project's registry on main; 'supported' means the architecture is registered, not that this checkpoint was tested.
Hardware fit
3 estimates
Hardware fit ESTIMATE
Weights-only memory for Bielik-11B-v3.0-Instruct: params × bytes/param × 1.2 overhead. KV cache and activations are extra. How this is computed.
| BF16 (2 bytes/param) | 27 GB · Single high-end GPU (<80 GB) |
|---|---|
| FP8 (1 byte/param) | 13 GB · Single GPU (<24 GB) |
| INT4 (0.5 bytes/param) | 6.7 GB · Single GPU (<24 GB) |
Weights only (params × bytes/param × 1.2 overhead); KV cache and activations extra.
All variants
30 variants
Lineage
Base model: speakleash/Bielik-11B-v3-Base-20250730
quantized: 20finetune: 6adapter: 4
Top variants (30 shown, ranked by downloads)
| Variant | Relation | Quant | Params | Band | License | Author | Likes | Downloads |
|---|---|---|---|---|---|---|---|---|
| speakleash/Bielik-11B-v3.0-Instruct-awq | quantized | awq | 11.3B | L | apache-2.0 | speakleash | 4 | 1.5M |
| speakleash/Bielik-11B-v3.0-Instruct-FP8-Dynamic | quantized | -FP8 | 11.2B | L | apache-2.0 | speakleash | 7 | 76k |
| speakleash/Bielik-11B-v3.0-Instruct-GGUF | quantized | GGUF | — | — | apache-2.0 | speakleash | 32 | 50.6k |
| bartowski/speakleash_Bielik-11B-v3.0-Instruct-GGUF | quantized | GGUF | — | — | — | bartowski | 5 | 10.8k |
| safestack/Bielik-11B-v3.0-Instruct-FP8-Dynamic | quantized | -FP8 | 11.2B | L | apache-2.0 | safestack | 0 | 6.6k |
| speakleash/Bielik-11B-v3.0-Instruct-MLX-8bit | quantized | MLX | 11.2B | L | apache-2.0 | speakleash | 3 | 3.4k |
| speakleash/Bielik-11B-v3.0-Instruct-MLX-4bit | quantized | MLX | 11.2B | L | apache-2.0 | speakleash | 2 | 3.4k |
| RomanMichtycz/Bielik-11B-v3-Instruct-Abliterated-Heretic-RP1 | quantized | GGUF | — | — | apache-2.0 | RomanMichtycz | 3 | 1.8k |
| LibraxisAI/Bielik-11B-v3.0-mlx-q4 | quantized | mlx | 11.2B | L | apache-2.0 | LibraxisAI | 1 | 1.7k |
| DevQuasar/speakleash.Bielik-11B-v3.0-Instruct-GGUF | quantized | GGUF | — | — | — | DevQuasar | 0 | 1.6k |
| gepardzik/Bielik-11B-v3.0-Instruct-heretic-MPOA | finetune | — | 11.2B | L | apache-2.0 | gepardzik | 2 | 1.3k |
| LibraxisAI/Bielik-11B-v3.0-mlx-bf16 | finetune | mlx | 11.2B | L | apache-2.0 | LibraxisAI | 0 | 1.1k |
| LibraxisAI/Bielik-11B-v3.0-mlx-q5 | quantized | mlx | 11.2B | L | apache-2.0 | LibraxisAI | 0 | 1.1k |
| norecyc/Bielik-11B-v3.0-Instruct-NVFP4 | quantized | NVFP4 | 6.40B | M | apache-2.0 | norecyc | 0 | 1k |
| LibraxisAI/Bielik-11B-v3.0-mlx-q8 | quantized | mlx | 11.2B | L | apache-2.0 | LibraxisAI | 0 | 922 |
| LibraxisAI/Bielik-11B-v3.0-mlx-mxfp4 | quantized | mlx | 11.2B | L | apache-2.0 | LibraxisAI | 0 | 849 |
| safestack/Bielik-11B-v3.0-Instruct-GGUF | quantized | GGUF | — | — | apache-2.0 | safestack | 1 | 681 |
| altomek/Bielik-11B-v3.0-Instruct-XXL-GGUF | quantized | GGUF | — | — | apache-2.0 | altomek | 1 | 566 |
| jmajkutewicz/Bielik-11B-v3.0-medadapt | finetune | — | 11.2B | L | other | jmajkutewicz | 1 | 540 |
| WMTH-100d2exam/Bielik-11B-v3.0-Instruct-matura-grpo-lora-closed-em | adapter | — | — | — | apache-2.0 | WMTH-100d2exam | 0 | 462 |
| directtt/Bielik-11B-v3.0-Instruct-Marych-FT | finetune | — | — | — | apache-2.0 | directtt | 0 | 443 |
| futurist-ai/Bielik-11B-v3.0-Instruct-MLX-4bit | quantized | MLX | 11.2B | L | apache-2.0 | futurist-ai | 1 | 298 |
| 14ai/Polando-HUGE-9B-GGUF | quantized | GGUF | — | — | apache-2.0 | 14ai | 1 | 104 |
| directtt/Bielik-11B-v3.0-Instruct-Marych-FT-v2 | finetune | — | — | — | apache-2.0 | directtt | 0 | 135 |
| Sticzu/bielik-nonsensik-v2 | adapter | GGUF | — | — | cc-by-nc-sa-4.0 | Sticzu | 0 | 128 |
| erybie222/bielik-supervised-fine-tuning-2026-08-29_10.11.34 | finetune | — | — | — | — | erybie222 | 2 | 0 |
| WMTH-100d2exam/Bielik-11B-v3.0-Instruct-matura-sft-lora-distill-epod | adapter | — | — | — | apache-2.0 | WMTH-100d2exam | 0 | 28 |
| jermalk/Bielik-11B-v3.0-Instruct-int4-ov | quantized | int4 | — | — | apache-2.0 | jermalk | 0 | 27 |
| WMTH-100d2exam/Bielik-11B-v3.0-Instruct-matura-sft-grpo-lora-closed-em | adapter | — | — | — | apache-2.0 | WMTH-100d2exam | 0 | 22 |
| jermalk/Bielik-11B-v3.0-Instruct-int8-ov | quantized | int8 | — | — | apache-2.0 | jermalk | 0 | 22 |
All providers
1 provider · 1 endpoint
Where to run 1 providers · 1 endpoints
Hosted endpoints observed for Bielik-11B-v3.0-Instruct, sorted by input price (USD per million tokens). Each row names its observer and date; prices change daily, so check the source before quoting. Cheapest observed input price: $0.40/M.
| Provider | Model ref | Context | Quant | $/M in | $/M out | tok/s | TTFT | Tools | Status | Observed |
|---|---|---|---|---|---|---|---|---|---|---|
| Public AI hf | speakleash/Bielik-11B-v3.0-Instruct | — | — | $0.40 | $0.40 | 51.1 | 1723 ms | yes | live | 2026-10-25 |
Papers
7 papers
Papers 7
arXiv links found in the model card and its tags. "Peer-reviewed" is shown only with evidence: a published venue in OpenAlex (measured) or the authors' own journal reference on arXiv (reported).
- Bielik v3 Small: Technical Report this model
- Bielik 11B v2 Technical Report this model
- HybridFlow: A Flexible and Efficient RLHF Framework cited in card
Similar models
3 models
Similar models
- SOLAR-10.7B-Instruct-v1.0 · 10.7B same size band (L), same task (text-generation)
- Llama-2-13b-chat-hf · 13.0B same size band (L), same task (text-generation), more providers
- SOLAR-10.7B-v1.0 · 10.7B same size band (L), same task (text-generation), permissive license
Sources
3 sources
Sources
Hub snapshot: 2026-10-25 · HF API: 2026-10-24 21:41 UTC · Providers: 2026-10-24 21:44 UTC · Articles: 2026-10-24 21:46 UTC · Summary: 2026-10-07 15:13 UTC · Last verified 2026-10-25
Follow Bielik-11B-v3.0-Instruct
Watch this model
- /feeds/models/speakleash--Bielik-11B-v3.0-Instruct.xml(RSS 2.0) · every change to Bielik-11B-v3.0-Instruct
- /feeds/families/speakleash--Bielik-v3.0.xml(RSS 2.0) · every change across the Bielik-v3.0 family
Email alerts are not open yet; the feeds above are the same data. See feeds for the item schema.