Task

Code

14 open-source models tuned for code.

671B-parameter MoE model with 37B active per token. Trained for roughly $5.6M of compute — a landmark in cost-efficient frontier training. Frontier-class quality at a fraction of the cost of the closed proprietary frontier. The DeepSeek licence permits commercial use with limited restrictions on military and unlawful applications. Running V3 yourself requires serious hardware (8× H100 at fp8); most teams will use it via the DeepSeek API or providers like Together.

Context: 128K
License: deepseek
VRAM Q4: 402.6 GB

DeepSeek R1

671B

Reasoning model trained with reinforcement learning on top of DeepSeek V3-Base. MIT licence — even the weights are unrestricted, making R1 the most permissively-licensed frontier reasoning model. Generates long internal chains-of-thought before answering, trading latency for accuracy on math, code, and reasoning benchmarks. Distilled variants (e.g. R1 Distill Llama 70B) recover most of the quality at much smaller scales.

Context: 128K
License: mit
VRAM Q4: 402.6 GB

DeepSeek Coder V2

236B

Coding-focused MoE model with 21B active parameters out of 236B total. Supports 338 programming languages with strong performance across mainstream stacks (Python, TypeScript, Go, Rust, Java, C++) and competent results on niche languages where most open models falter. The DeepSeek licence applies — commercial use permitted with some application restrictions.

Context: 128K
License: deepseek
VRAM Q4: 141.6 GB

Mixtral 8×22B Instruct

141B

Scaled-up Mixtral with 22B-parameter experts. ~39B active parameters out of 141B total. Strong long-context performance and competitive coding scores. Apache 2.0 makes it attractive for self-hosting where the licence terms of Llama 3 are a non-starter.

Context: 66K
License: apache-2-0
VRAM Q4: 84.6 GB

Qwen2.5 72B Instruct

72B

The flagship Qwen 2.5 release. Competes with Llama 3.1 405B on many benchmarks at one-fifth the parameter count. Note the 72B specifically uses the Qwen License (commercial use up to 100M MAU) — the smaller Qwen2.5 sizes are Apache 2.0.

Context: 128K
License: qwen
VRAM Q4: 43.2 GB

Mixtral 8×7B Instruct

46.7B

The mixture-of-experts release that introduced 8 experts of 7B each, 2 active per token. ~13B active parameters with 47B total, which makes per-token inference roughly as fast as a 13B dense model while approaching 70B dense quality. Apache 2.0 weights mean it's still a popular self-hosting choice. Memory footprint is the main constraint — the full 47B parameters must be loaded even though only a quarter are active per token.

Context: 33K
License: apache-2-0
VRAM Q4: 28 GB

Qwen2.5 32B Instruct

32B

32B sweet-spot model: strong reasoning, fits on one H100 in fp16, on a 4090 at Q4. The 32B size in particular hits a quality/cost knee — quality scales with parameters faster than cost up to ~32B, and slower afterwards. Favoured for production chat where 7B isn't sharp enough and where 70B+ would over-spec the hardware budget. Apache 2.0 licence.

Context: 128K
License: apache-2-0
VRAM Q4: 19.2 GB

Qwen2.5 Coder 32B

32B

Coding-specialised Qwen2.5 32B fine-tune. GPT-4o-class on HumanEval and BigCodeBench at the time of release. Trained on additional code-heavy data with extended pre-training. Apache 2.0. Natural pick for self-hosted coding assistants, code-review automation, and any agent loop that primarily writes code.

Context: 128K
License: apache-2-0
VRAM Q4: 19.2 GB

Mistral Small 3

24B

24B dense model from Mistral's January 2025 release that competes with Llama 3.3 70B on many tasks at a third of the parameter count. Apache 2.0 licensed and small enough to run on a single 4090 at Q4. Good pick when you want Llama-3.3-70B-class chat quality but at a friendlier hardware budget, or when the licence matters and Llama's community terms don't fit.

Context: 33K
License: apache-2-0
VRAM Q4: 14.4 GB

Phi-4 14B

14B

14B model trained primarily on synthetic data. Punches above its weight on reasoning, especially MATH and GPQA. MIT licensed. A standout choice when you want strong reasoning quality without paying 70B-tier hardware costs. Phi-4 in particular demonstrated that careful synthetic-data curation can extract frontier-class reasoning from a relatively small dense model.

Context: 16K
License: mit
VRAM Q4: 8.4 GB

Qwen2.5 14B Instruct

14B

Mid-size Qwen2.5 with broad task coverage. The sweet spot for users who want noticeably better quality than 7B but can't justify the hardware footprint of 32B or 72B.

Context: 128K
License: apache-2-0
VRAM Q4: 8.4 GB

Mistral Nemo 12B

12B

Joint Mistral × NVIDIA model with 128K context, designed as a drop-in upgrade to Mistral 7B. Trained with NVIDIA's Megatron stack and released under Apache 2.0. Strong multilingual coverage thanks to the Tekken tokenizer.

Context: 128K
License: apache-2-0
VRAM Q4: 7.2 GB

Llama 3.1 8B Instruct

The workhorse 8B instruction-tuned model. Excellent quality-to-cost ratio and the broadest ecosystem support of any open-weights model — every major inference engine, fine-tuning library, and quantization toolchain has a 3.1 8B preset. Fits in 24 GB of VRAM at fp16, ~6 GB at Q4. Strong default for production chat where 70B is overkill, for fine-tuning on a specialist task, and for any workload where you want a known-good baseline.

Context: 128K
License: llama-3
VRAM Q4: 4.8 GB

Qwen2.5 7B Instruct

Apache-2.0-licensed 7B model with surprisingly strong reasoning and multilingual chops. Qwen 2.5 trains on a larger and more carefully filtered corpus than the original Qwen series, and the 7B variant punches well above its weight on coding and math benchmarks. A strong default for cost-sensitive chat workloads and for fine-tuning experiments where the Apache licence simplifies downstream redistribution.

Context: 128K
License: apache-2-0
VRAM Q4: 4.2 GB