gpt-oss-20b vs Qwen3 30B A3B for local chat
Two mixture-of-experts chat models for a 16 to 32 GB machine
Facts from the reviewed catalog
| Specification | gpt-oss-20b | Qwen3 30B A3B Instruct 2507 |
|---|---|---|
| Task | Text & chat | Text & chat |
| Parameters | 21B · 3.6B active | 30.5B · 3.3B active |
| License | Apache 2.0 | Apache 2.0 |
| Commercial use | Allowedlicense text | Allowedlicense text |
| Quality · LMArena 2026-10-22 · text | 1,317 ± 61,311–1,324 · 10,417 votes · rank 269 · source | 1,382 ± 51,377–1,387 · 23,306 votes · rank 190 · source |
| Memory | Q4 planning estimate (~16 GB of weights before the machine's reserve) | Q4 planning estimate (~23 GB of weights before the machine's reserve) |
| Checkpoint | official card | official card |
In LMArena 2026-10-22 · text, Qwen3 30B A3B Instruct 2507 rates 1,382 ± 5 against 1,317 ± 6 for gpt-oss-20b, 65 points ahead with intervals that do not overlap. Blind pairwise votes by people on the hosted model, CC BY 4.0. A rating is relative; overlapping intervals mean the order is not settled. Scale: Arena rating (style control, overall) ± half the confidence interval.
Memory on three machines
Each machine's own reserve is added (graphics cards 2 GB; unified and system memory 25%, 2 to 8 GB). Planning estimates, not measured runs.
| Machine | gpt-oss-20b | Qwen3 30B A3B Instruct 2507 |
|---|---|---|
| RTX 4060 Ti · 16 GB VRAM | ~18 GBIn memory, outside the fast part · offload to PC RAM (32 GB assumed) | ~25 GBIn memory, outside the fast part · offload to PC RAM (32 GB assumed) |
| RTX 4090 · 24 GB VRAM | ~18 GBFast memory holds it | ~25 GBIn memory, outside the fast part · offload to PC RAM (32 GB assumed) |
| MacBook Air 13-inch M4 · 32 GB unified | ~24 GBFast memory holds it | ~31 GBFast memory holds it |
gpt-oss-20b: Q4 planning estimate (~16 GB of weights before the machine's reserve). Qwen3 30B A3B Instruct 2507: Q4 planning estimate (~23 GB of weights before the machine's reserve). Formulas
Everything that runs on each machine: RTX 4060 Ti · 16 GB VRAM · RTX 4090 · 24 GB VRAM · MacBook Air 13-inch M4 · 32 GB unified
Which to choose
Both are mixtures of experts with a small active part (3.6 billion parameters for gpt-oss-20b, 3.3 billion for Qwen3 30B A3B), so both generate quickly for their size. gpt-oss-20b is smaller in total (21 billion against 30.5) and OpenAI ships it in 4-bit MXFP4, stating it runs within 16 GB; the generic Q4 estimate below is more cautious than that statement. Qwen3 30B A3B spills past a 24 GB card at Q4 and fits a 32 GB Mac. Both are Apache 2.0. Qwen3 30B A3B publishes a 262,144-token context against 131,072, and rates higher in the same LMArena text release.
Files and run commands
The default file of the most downloaded collected source, or the commands each card documents. Not run by this site.
gpt-oss-20b
unsloth/gpt-oss-20b-GGUF · gpt-oss-20b-Q4_K_M.gguf · Q4_K_M · 11.6 GB · unsloth
ollama run hf.co/unsloth/gpt-oss-20b-GGUF:Q4_K_Mllama-server -hf unsloth/gpt-oss-20b-GGUF:Q4_K_Mlms get https://huggingface.co/unsloth/gpt-oss-20b-GGUF- mlx-lm: mlx-lm runs MLX repositories; pick the MLX version on a Mac
Qwen3 30B A3B Instruct 2507
unsloth/Qwen3-30B-A3B-Instruct-2507-GGUF · Qwen3-30B-A3B-Instruct-2507-Q4_K_M.gguf · Q4_K_M · 18.6 GB · unsloth
ollama run hf.co/unsloth/Qwen3-30B-A3B-Instruct-2507-GGUF:Q4_K_Mllama-server -hf unsloth/Qwen3-30B-A3B-Instruct-2507-GGUF:Q4_K_Mlms get https://huggingface.co/unsloth/Qwen3-30B-A3B-Instruct-2507-GGUF- mlx-lm: mlx-lm runs MLX repositories; pick the MLX version on a Mac
Numbers each publisher reports
Each card's own results with its own settings. Publishers choose different tests and settings, so these values are never compared across publishers here.
gpt-oss-20b · reported by OpenAI
Qwen3 30B A3B Instruct 2507 · reported by Alibaba
- MMLU-Pro
- 78.4Knowledge
- MMLU-Redux
- 89.3Knowledge
- GPQA
- 70.4Knowledge
- SuperGPQA
- 53.4Knowledge
- AIME25
- 61.3Reasoning
- HMMT25
- 43.0Reasoning
- ZebraLogic
- 90.0Reasoning
- LiveBench 20241125
- 69.0Reasoning
Every reported result
- LiveCodeBench v6 (25.02-25.05)
- 43.2Coding
- MultiPL-E
- 83.8Coding
- Aider-Polyglot
- 35.6Coding
- IFEval
- 84.7Alignment
- Arena-Hard v2
- 69.0Alignment
- Creative Writing v3
- 86.0Alignment
- WritingBench
- 85.5Alignment
- BFCL-v3
- 65.1Agent
- TAU1-Retail
- 59.1Agent
- TAU1-Airline
- 40.0Agent
- TAU2-Retail
- 57.0Agent
- TAU2-Airline
- 38.0Agent
- TAU2-Telecom
- 12.3Agent
- MultiIF
- 67.9Multilingualism
- MMLU-ProX
- 72.0Multilingualism
- INCLUDE
- 71.9Multilingualism
- PolyMATH
- 43.1Multilingualism
Values as printed on the card, with the publisher's own settings (card section: Performance); column “Qwen3-30B-A3B-Instruct-2507”. Not comparable across publishers and not used in any ranking. Model card
Limits and dates
The published quality benchmark does not test a local quantized build. Memory estimates are not speed measurements or guaranteed hardware fits. Repository size, when available, is not VRAM. Formulas
Page updated · Catalog reviewed · LMArena 2026-10-22 · text