The story so far
Weights, every dated article, and repo commits on one rail.
8 repo commits since the weights landed
Teal is the weights (2026-06-05), grey ticks are repo commits.
North-Mini-Code-1.0
CohereLabs Apache 2.0 out 5 months ago
30.5B parameters (8 of 128 experts per token), text in, text out, 488K tokens of context. 1 provider serves it.
Verified 2026-10-25 Rank #626 in LLMs · score 73.4 method v1 Sources 4 Downloads 30d 7.3k · #525 of 619 tracked
Do people use it?7.3kdownloads, 30 days#525 of 619 we track · below the median+2 likes, downloads −3% this week583 likesHugging Face2026-10-24
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 | Cohere · free tier | $0 | $0 | 250K | unknown | Oct 24 |
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; 39 quantized builds for local use.
- As a service1Cohere
- On own hardware39quantized builds
- Built on44 fine-tunes · 0 adapters · 0 merges
- In apps14Hugging Face Spaces
Top Spaces running it: CohereLabs/North-Mini-Code-1.0 ♥ 47 · build-small-hackathon/tiny-army ♥ 5
Will it fit?
Weights only, an estimate: leave room for the KV cache. Method
Fits a 64 GB Mac (unified memory) at INT4
About 18 GB at INT4, 37 GB at FP8, 73 GB at BF16: weights only, so leave room for the KV cache. Every expert must be resident, so total parameters count.
Estimate params × bytes per param × 1.2 · how we estimate
What people built on it
43 derivatives on the Hub. Which build to run, by runtime: the most downloaded in each class.
43 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.
Benchmarks
3 rows
Benchmarks 3 rows
measured third-party leaderboard rowauthor-reported number from the model card or eval files (3)No confidence intervals unless the source gave them; settings differ between rows, so compare within a benchmark, not across. Date = the benchmark's own date when it gives one, else when we read it.
| Benchmark | Value | Setting / notes | Reported by | Date |
|---|---|---|---|---|
| author-reported SWE_Bench_Pro (ScaleAI/SWE-bench_Pro) | 40.2 | — | North Mini Code Model Page | 2026-06-09 |
| author-reported swe_bench_%_resolved (SWE-bench/SWE-bench_Verified) | 67.6 | — | North Mini Code Model Page | 2026-06-09 |
| author-reported terminalbench_2 (harborframework/terminal-bench-2.0) | 36 | — | North Mini Code Model Page | 2026-06-09 |
What the card says
2 cited sentences
Summary generated, cited
North Mini Code is a 30.5B parameter mixture-of-experts model optimized for code generation, agentic software engineering, and terminal tasks.[1][2] The architecture utilizes 128 experts with 8 activated per token and supports a context length of 256K.[3][4][5]
- parameters human
- “optimized for code generation, agentic software engineeri…”
- num experts
- experts per token
- “context length of 256K”
Generated 2026-10-07 15:12 UTC by parasail-gemma-4-26b-a4b-it; each numeral points at the fact it rests on. Not a verdict.
Use-case evidence
6 use cases
What it is good at 6 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.
- Write code with agentsauthor-reportedAuthor reports SWE_Bench_Pro 40.2Author reports swe_bench_%_resolved 67.6Author reports terminalbench_2 36.0+3 more
- Call tools and APIsstructuraltag: agentchat template: tool_calls
- Chat and follow instructionsstructuraltag: conversationalchat template present
- Read long documentsstructuraldecision-model estimate p=0.98config: 500,000 context
- Reason through hard problemsstructuralchat template: reasoning_content
- Run on a laptop or phonestructuralname: Mini
Facts
21 fields
Facts
- North-Mini-Code-1.0 has 30.5B parameters. (safetensors metadata, observed 2026-10-24)
- North-Mini-Code-1.0 is a mixture-of-experts model with 128 experts, routing 8 per token. (config.json, observed 2026-10-24)
- Context length is 500,000 tokens (488K). (config.json max_position_embeddings, observed 2026-10-24)
- License: Apache 2.0 (friction: none). Permissive; patent grant. (license file, observed 2026-10-24)
- Created 2026-06-05, last modified 2026-06-15. (HF API, observed 2026-10-24)
| Pipeline tag | text-generation |
|---|---|
| Library | transformers |
| Architectures | Cohere2MoeForCausalLM |
| Model type | cohere2_moe |
| Dtype | bfloat16 |
| Params by dtype | BF16: 30,484,303,872 |
| Size band | L |
| MoE | yes |
| Experts | 128 |
| Experts per token | 8 |
| Context length | 500,000 (config.json max_position_embeddings) |
| License | Apache 2.0 (apache-2.0) |
| Hub snapshot | 2026-10-25 |
Servability
vLLM · SGLang
Servability
vLLM: supported · Cohere2MoeForCausalLM · 7050ef881ff0835f6890a47db1859a24c65481d6SGLang: supported · Cohere2MoeForCausalLM · 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 North-Mini-Code-1.0: params × bytes/param × 1.2 overhead. KV cache and activations are extra. How this is computed.
| BF16 (2 bytes/param) | 73 GB · Single high-end GPU (<80 GB) |
|---|---|
| FP8 (1 byte/param) | 37 GB · Single high-end GPU (<80 GB) |
| INT4 (0.5 bytes/param) | 18 GB · Single GPU (<24 GB) |
Weights only (params × bytes/param × 1.2 overhead); KV cache and activations extra. MoE: all experts must be resident, so total parameters count.
All variants
40 variants
Lineage
quantized: 39finetune: 4
Top variants (40 shown, ranked by downloads)
| Variant | Relation | Quant | Params | Band | License | Author | Likes | Downloads |
|---|---|---|---|---|---|---|---|---|
| unsloth/North-Mini-Code-1.0-GGUF | quantized | GGUF | — | — | apache-2.0 | unsloth | 146 | 750.3k |
| bartowski/North-Mini-Code-1.0-GGUF | quantized | GGUF | — | — | apache-2.0 | bartowski | 5 | 66.7k |
| CohereLabs/North-Mini-Code-1.0-fp8 | quantized | -fp8 | 30.5B | L | apache-2.0 | CohereLabs | 32 | 52.5k |
| CohereLabs/North-Mini-Code-1.0-w4a16 | quantized | -w4a16 | 30.5B | L | apache-2.0 | CohereLabs | 61 | 16.5k |
| mradermacher/North-Mini-Code-1.0-i1-GGUF | quantized | GGUF | — | — | apache-2.0 | mradermacher | 2 | 16.5k |
| Arki05/North-Mini-Code-1.0-GGUF | quantized | GGUF | — | — | apache-2.0 | Arki05 | 9 | 7.9k |
| XanuNetworks/North-Mini-Code-1.0-NVFP4 | quantized | NVFP4 | 17.4B | L | apache-2.0 | XanuNetworks | 1 | 5.4k |
| byteshape/North-Mini-Code-1.0-GGUF | quantized | GGUF | — | — | apache-2.0 | byteshape | 8 | 4.8k |
| mlx-community/North-Mini-Code-1.0-4bit | quantized | -4bit | 30.5B | L | apache-2.0 | mlx-community | 3 | 5k |
| cyankiwi/North-Mini-Code-1.0-AWQ-INT4 | quantized | AWQ | 30.5B | L | apache-2.0 | cyankiwi | 1 | 3k |
| DevQuasar/CohereLabs.North-Mini-Code-1.0-GGUF | quantized | GGUF | — | — | — | DevQuasar | 0 | 3k |
| maxkru92/North-Mini-Code-1.0-GGUF | quantized | GGUF | — | — | apache-2.0 | maxkru92 | 0 | 2.9k |
| brodatech/North-Mini-Code-1.0-GGUF | quantized | GGUF | — | — | apache-2.0 | brodatech | 0 | 2.7k |
| mlx-community/North-Mini-Code-1.0-6bit | quantized | — | 30.5B | L | apache-2.0 | mlx-community | 3 | 2.4k |
| bsisduck/North-Mini-Code-1.0-MLX-MXFP4 | quantized | MLX | 30.5B | L | apache-2.0 | bsisduck | 3 | 2.3k |
| mradermacher/North-Mini-Code-1.0-GGUF | quantized | GGUF | — | — | apache-2.0 | mradermacher | 0 | 2.2k |
| mlx-community/North-Mini-Code-1.0-bf16 | finetune | -bf16 | 30.5B | L | apache-2.0 | mlx-community | 3 | 1.2k |
| unsloth/North-Mini-Code-1.0 | finetune | — | 30.5B | L | apache-2.0 | unsloth | 8 | 983 |
| heimann/North-Mini-Code-1.0-GGUF | quantized | GGUF | — | — | apache-2.0 | heimann | 2 | 1.2k |
| michaelw9999/North-Mini-Code-1.0-GGUF | quantized | GGUF | — | — | — | michaelw9999 | 3 | 1.1k |
| deucebucket/North-Mini-Code-Cerebellum-GGUF | quantized | GGUF | — | — | cc-by-nc-4.0 | deucebucket | 2 | 1.1k |
| ji-farthing/North-Mini-Code-1.0-ik-llama-validation-GGUF | quantized | GGUF | — | — | apache-2.0 | ji-farthing | 1 | 1k |
| mlx-community/North-Mini-Code-1.0-5bit | quantized | — | 30.5B | L | apache-2.0 | mlx-community | 1 | 1k |
| bsisduck/North-Mini-Code-1.0-MLX-MXFP8 | quantized | MLX | 30.5B | L | apache-2.0 | bsisduck | 1 | 1k |
| FreedomAISVR/North-Mini-Code-1.0-NVFP4-GGUF | quantized | NVFP4 | — | — | apache-2.0 | FreedomAISVR | 0 | 1k |
| FreedomAISVR/North-Mini-Code-1.0-MXFP4-GGUF | quantized | MXFP4 | — | — | apache-2.0 | FreedomAISVR | 0 | 846 |
| KellHect/Borealis-Code-1.0 | quantized | GGUF | 30.5B | L | apache-2.0 | KellHect | 1 | 724 |
| kingjones777/North-Mini-Code-1.0-ROCmFP4-STRIX-GGUF | quantized | GGUF | — | — | apache-2.0 | kingjones777 | 0 | 695 |
| sahilchachra/north-mini-code-mxfp4-mlx | quantized | mxfp4 | 30.5B | L | apache-2.0 | sahilchachra | 1 | 600 |
| NANI-Nithin/north-mini-code-gguf | quantized | gguf | — | — | apache-2.0 | NANI-Nithin | 1 | 597 |
| bsisduck/North-Mini-Code-1.0-MLX-BF16 | finetune | MLX | 30.5B | L | apache-2.0 | bsisduck | 1 | 582 |
| theroadnotbacon/BLS-Mini-Code-1.0-Q4-MLX | quantized | MLX | 30.5B | L | — | theroadnotbacon | 2 | 396 |
| AdmiralGloom/North-Mini-Code-1.0-GGUF-Q6 | quantized | GGUF | — | — | apache-2.0 | AdmiralGloom | 1 | 444 |
| FenomAI/North-Mini-Code-1.0-AWQ-INT4 | quantized | AWQ | 32.4B | L | apache-2.0 | FenomAI | 0 | 484 |
| aaronmwelborn/North-Mini-Code-1.0-8bit | quantized | -8bit | 30.5B | L | apache-2.0 | aaronmwelborn | 0 | 424 |
| LLMWildling/North-Mini-Code-1.0-43B-a5b | quantized | — | 40.7B | XL | apache-2.0 | LLMWildling | 2 | 219 |
| Fr4nkSt3in/North-Mini-Code-1.0-BF16-GGUF | quantized | -BF16 | — | — | apache-2.0 | Fr4nkSt3in | 0 | 248 |
| sakamakismile/BLS-Mini-Code-1.0-NVFP4 | quantized | NVFP4 | 17.4B | L | cc-by-nc-4.0 | sakamakismile | 1 | 170 |
| NarsAI/Mini-Code-1.7 | quantized | — | 30.5B | L | apache-2.0 | NarsAI | 0 | 184 |
| AlexWortega/North-Mini-Code-1.0-EAGLE3 | finetune | — | 183M | S | cc-by-nc-4.0 | AlexWortega | 1 | 104 |
Repository changelog
8 commits
Repository changelog 8 commits
| Date | Commit | SHA |
|---|---|---|
| 2026-06-15 | Remove outdated and broken chat templates (#9) | d11e61a |
| 2026-06-11 | Add eval results for SWE-bench Verified, SWE-bench Pro, and Terminal-Bench v2 (#7) | 7cbbd56 |
| 2026-06-09 | Update README.md | effaeda |
| 2026-06-09 | Create README.md | 98cd612 |
| 2026-06-09 | Update chat_template.jinja | 85b92fb |
| 2026-06-08 | fix interleaved reasoning | 586cec0 |
| 2026-06-06 | Update tokenizer JSON | 8e2b162 |
| 2026-06-05 | Initial commit | 82a8417 |
All providers
1 provider · 1 endpoint
Where to run 1 providers · 1 endpoints
Hosted endpoints observed for North-Mini-Code-1.0, 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: $0.00/M.
| Provider | Model ref | Context | Quant | $/M in | $/M out | tok/s | TTFT | Tools | Status | Observed |
|---|---|---|---|---|---|---|---|---|---|---|
| Cohere openrouter | cohere/north-mini-code:free | 250K | unknown | $0.00 | $0.00 | — | — | yes | live | 2026-10-24 |
Similar models
3 models
Similar models
- GLM-4.7-Flash · 31.2B same size band (L), also MoE, same task (text-generation), more providers, permissive license
- Qwen3-Coder-30B-A3B-Instruct · 30.5B same size band (L), also MoE, same task (text-generation), more providers, permissive license
- Qwen3-30B-A3B · 30.5B same size band (L), also MoE, same task (text-generation), more providers, permissive license
Sources
4 sources
Sources
Hub snapshot: 2026-10-25 · HF API: 2026-10-24 21:40 UTC · Providers: 2026-10-24 21:44 UTC · Articles: 2026-10-24 21:46 UTC · Summary: 2026-10-07 15:12 UTC · Last verified 2026-10-25
Follow North-Mini-Code-1.0
Watch this model
- /feeds/models/CohereLabs--North-Mini-Code-1.0.xml(RSS 2.0) · every change to North-Mini-Code-1.0
- /feeds/families/CohereLabs--North-1.0.xml(RSS 2.0) · every change across the North-1.0 family
Email alerts are not open yet; the feeds above are the same data. See feeds for the item schema.