Seed OSS 36B Instruct vs Qwen3 30B A3B Thinking 2507
Seed OSS 36B Instruct and Qwen3 30B A3B Thinking 2507 match in llmrun Score, with values of 102.5 versus 100.6; both fit 24 GB at Q4_K_M; Qwen3 30B A3B Thinking 2507 is 6.0× faster on RTX 4090 (~173 vs ~29 tok/s).
Model A · ByteDance-Seed
Seed OSS 36B Instruct36.2B · 512K context · Apache 2.0 · Aug 2025
Model B · Alibaba
Qwen3 30B A3B Thinking 250730.5B · 3.4B active · 256K context · Apache 2.0 · Jul 2025
Choose Seed OSS 36B Instruct if you want…
Ahead on the benchmarks
3 of 4 shared boards
Longer context window
512K vs 256K tokens
Choose Qwen3 30B A3B Thinking 2507 if you want…
Fits a 16 GB GPU
Q3_K_M · 16 GB; Seed OSS 36B Instruct needs 12.4 GB at its smallest quant
6.0× faster on RTX 4090
~173 vs ~29 tok/s
Lighter on memory
18.7 vs 22.3 GB at Q4_K_M
On your hardware
Each model at its default quant (Q4_K_M), a smaller one when it must. Estimates, ±20%.
8 GB
RTX 4060 · 8 GB tier
Seed OSS 36B Instruct
Doesn't fit
smallest quant needs 12.4 GB
Qwen3 30B A3B Thinking 2507
Doesn't fit
smallest quant needs 9.5 GB
12 GB
RTX 3060 12GB · 12 GB tier
Seed OSS 36B Instruct
Doesn't fit
smallest quant needs 12.4 GB
Qwen3 30B A3B Thinking 2507
IQ2_M · ~130 tok/s
11.4 GB · up to 16Kfits
16 GB
RTX 5060 Ti 16GB · 16 GB tier
Seed OSS 36B Instruct
IQ2_M · ~23 tok/s
14.7 GB · up to 16K
Qwen3 30B A3B Thinking 2507
Q3_K_M · ~119 tok/s
16 GB · up to 16Kbetter quant
24 GBreference
RTX 4090 · 24 GB tier
Seed OSS 36B Instruct
Q4_K_M · ~29 tok/stight
24.1 GB · up to 8K
Qwen3 30B A3B Thinking 2507
Q4_K_M · ~173 tok/s
19.4 GB · up to 64Kmore headroom
32 GB
RTX 5090 · 32 GB tier
Seed OSS 36B Instruct
Q4_K_M · ~52 tok/s
24.1 GB · up to 64K
Qwen3 30B A3B Thinking 2507
Q4_K_M · ~227 tok/s
19.4 GB · up to 256Kfaster
48 GB
RTX A6000 · 48 GB tier
Seed OSS 36B Instruct
Q4_K_M · ~22 tok/s
24.1 GB · up to 128K
Qwen3 30B A3B Thinking 2507
Q4_K_M · ~148 tok/s
19.4 GB · up to 256Kfaster
Mac 32 GB
Mac Mini M4 · Macs
Seed OSS 36B Instruct
Q4_K_M · ~4 tok/s
24.1 GB · up to 32K
Qwen3 30B A3B Thinking 2507
Q4_K_M · ~29 tok/s
19.4 GB · up to 128Kfaster
Mac 64 GB
Mac Studio M4 Max · Macs
Seed OSS 36B Instruct
Q4_K_M · ~17 tok/s
24.1 GB · up to 256K
Qwen3 30B A3B Thinking 2507
Q4_K_M · ~67 tok/s
19.4 GB · up to 256Kfaster
Mac 128 GB
Mac Studio M4 Max · Macs
Seed OSS 36B Instruct
Q4_K_M · ~17 tok/s
24.1 GB · up to 512K
Qwen3 30B A3B Thinking 2507
Q4_K_M · ~67 tok/s
19.4 GB · up to 256Kfaster
Reference GPU: RTX 4090. Tinted cell = the better pick on that machine (fits, more headroom, or at least 10% faster).
Overall
Seed OSS 36B Instruct
Qwen3 30B A3B Thinking 2507
Science
Seed OSS 36B Instruct
no data
Qwen3 30B A3B Thinking 2507
Reasoning
Seed OSS 36B Instruct
no data
Qwen3 30B A3B Thinking 2507
Seed OSS 36B Instruct · #42 of 92 open models · #8 of 34 that fit 24 GB
Qwen3 30B A3B Thinking 2507 · #45 of 92 open models · #9 of 34 that fit 24 GB
Benchmarks
4 shared boards: Seed OSS 36B Instruct leads 3, Qwen3 30B A3B Thinking 2507 leads 1 · Benchmarks imported Oct 2026
Seed OSS 36B Instruct
71.5%#30/97 open
IndependentEpoch AI · Aug 2025
Qwen3 30B A3B Thinking 2507
70.1%#32/97 open
IndependentEpoch AI · Jul 2025
Seed OSS 36B Instruct
67.5%#29/85 open
IndependentEpoch AI · Aug 2025
Qwen3 30B A3B Thinking 2507
70.3%#27/85 open
IndependentEpoch AI · Jul 2025
Seed OSS 36B Instruct
82.7%#15/144 open
IndependentTIGER-Lab
Qwen3 30B A3B Thinking 2507
80.9%#22/144 open
IndependentTIGER-Lab
Seed OSS 36B Instruct
13.0%#18/67 open
IndependentEpoch AI · Aug 2025
Qwen3 30B A3B Thinking 2507
8.0%#26/67 open
IndependentEpoch AI · Jul 2025
VRAM by quantization
Memory to load the model and 2K context; below it, the total at 32K and 128K context. Lighter side in bold.
Q2_K
Seed OSS 36B Instruct
15.9 GB13.6 GB file
32K 20.0 GB · 128K 32.8 GB
Qwen3 30B A3B Thinking 2507
13.4 GB11.3 GB file
32K 14.9 GB · 128K 19.7 GB
Q3_K_S
Seed OSS 36B Instruct
16.4 GB15.9 GB file
32K 20.4 GB · 128K 33.3 GB
Qwen3 30B A3B Thinking 2507
13.8 GB13.3 GB file
32K 15.3 GB · 128K 20.1 GB
Q3_K_M
Seed OSS 36B Instruct
18.2 GB17.6 GB file
32K 22.2 GB · 128K 35.1 GB
Qwen3 30B A3B Thinking 2507
15.3 GB14.7 GB file
32K 16.8 GB · 128K 21.6 GB
Q4_0
Seed OSS 36B Instruct
18.6 GB20.6 GB file
32K 22.7 GB · 128K 35.6 GB
Qwen3 30B A3B Thinking 2507
15.7 GB17.4 GB file
32K 17.2 GB · 128K 22.0 GB
Q4_K_M
Seed OSS 36B Instruct
22.3 GB21.8 GB file
32K 26.3 GB · 128K 39.2 GB
Qwen3 30B A3B Thinking 2507
18.7 GB18.6 GB file
32K 20.2 GB · 128K 25.1 GB
Q5_K_M
Seed OSS 36B Instruct
26.3 GB25.6 GB file
32K 30.4 GB · 128K 43.2 GB
Qwen3 30B A3B Thinking 2507
22.2 GB21.7 GB file
32K 23.7 GB · 128K 28.5 GB
Q6_K
Seed OSS 36B Instruct
30.4 GB29.7 GB file
32K 34.4 GB · 128K 47.3 GB
Qwen3 30B A3B Thinking 2507
25.6 GB25.1 GB file
32K 27.1 GB · 128K 31.9 GB
Q8_0
Seed OSS 36B Instruct
36.7 GB38.4 GB file
32K 40.8 GB · 128K 53.6 GB
Qwen3 30B A3B Thinking 2507
30.9 GB32.5 GB file
32K 32.4 GB · 128K 37.3 GB
19 more quants
IQ2_XXS
Seed OSS 36B Instruct
10.5 GB10.2 GB file
32K 14.5 GB · 128K 27.4 GB
Qwen3 30B A3B Thinking 2507
8.8 GB10.3 GB file
32K 10.3 GB · 128K 15.1 GB
IQ2_XS
Seed OSS 36B Instruct
11.4 GB10.9 GB file
32K 15.4 GB · 128K 28.3 GB
Qwen3 30B A3B Thinking 2507
9.6 GB8.7 GB file
32K 11.1 GB · 128K 15.9 GB
IQ2_S
Seed OSS 36B Instruct
11.9 GB11.6 GB file
32K 15.9 GB · 128K 28.8 GB
Qwen3 30B A3B Thinking 2507
9.9 GB8.7 GB file
32K 11.5 GB · 128K 16.3 GB
IQ2_M
Seed OSS 36B Instruct
12.8 GB12.7 GB file
32K 16.8 GB · 128K 29.7 GB
Qwen3 30B A3B Thinking 2507
10.7 GB10.8 GB file
32K 12.2 GB · 128K 17.1 GB
IQ3_XXS
Seed OSS 36B Instruct
14.6 GB14.3 GB file
32K 18.6 GB · 128K 31.5 GB
Qwen3 30B A3B Thinking 2507
12.2 GB12.9 GB file
32K 13.7 GB · 128K 18.6 GB
Q2_K_S
Seed OSS 36B Instruct
15.0 GB12.6 GB file
32K 19.1 GB · 128K 31.9 GB
Qwen3 30B A3B Thinking 2507
12.6 GB10.5 GB file
32K 14.1 GB · 128K 19.0 GB
IQ3_XS
Seed OSS 36B Instruct
15.5 GB15.1 GB file
32K 19.5 GB · 128K 32.4 GB
Qwen3 30B A3B Thinking 2507
13.0 GB12.7 GB file
32K 14.5 GB · 128K 19.3 GB
IQ3_S
Seed OSS 36B Instruct
15.9 GB15.9 GB file
32K 20.0 GB · 128K 32.8 GB
Qwen3 30B A3B Thinking 2507
13.4 GB13.3 GB file
32K 14.9 GB · 128K 19.7 GB
IQ3_M
Seed OSS 36B Instruct
16.8 GB16.5 GB file
32K 20.9 GB · 128K 33.8 GB
Qwen3 30B A3B Thinking 2507
14.1 GB14.1 GB file
32K 15.7 GB · 128K 20.5 GB
Q3_K_L
Seed OSS 36B Instruct
19.1 GB19.1 GB file
32K 23.1 GB · 128K 36.0 GB
Qwen3 30B A3B Thinking 2507
16.1 GB14.6 GB file
32K 17.6 GB · 128K 22.4 GB
IQ4_XS
Seed OSS 36B Instruct
20.0 GB19.5 GB file
32K 24.0 GB · 128K 36.9 GB
Qwen3 30B A3B Thinking 2507
16.8 GB16.4 GB file
32K 18.3 GB · 128K 23.2 GB
IQ4_NL
Seed OSS 36B Instruct
20.9 GB20.6 GB file
32K 24.9 GB · 128K 37.8 GB
Qwen3 30B A3B Thinking 2507
17.6 GB17.3 GB file
32K 19.1 GB · 128K 23.9 GB
Q4_1
Seed OSS 36B Instruct
20.9 GB22.8 GB file
32K 24.9 GB · 128K 37.8 GB
Qwen3 30B A3B Thinking 2507
17.6 GB19.2 GB file
32K 19.1 GB · 128K 23.9 GB
Q4_K_S
Seed OSS 36B Instruct
20.9 GB20.7 GB file
32K 24.9 GB · 128K 37.8 GB
Qwen3 30B A3B Thinking 2507
17.6 GB17.5 GB file
32K 19.1 GB · 128K 23.9 GB
Q4_K_L
Seed OSS 36B Instruct
22.7 GB22.4 GB file
32K 26.7 GB · 128K 39.6 GB
Qwen3 30B A3B Thinking 2507
19.1 GB18.9 GB file
32K 20.6 GB · 128K 25.4 GB
Q5_0
Seed OSS 36B Instruct
23.2 GB25.0 GB file
32K 27.2 GB · 128K 40.1 GB
Qwen3 30B A3B Thinking 2507
19.5 GB21.1 GB file
32K 21.0 GB · 128K 25.8 GB
Q5_K_S
Seed OSS 36B Instruct
25.4 GB25.0 GB file
32K 29.5 GB · 128K 42.3 GB
Qwen3 30B A3B Thinking 2507
21.4 GB21.1 GB file
32K 22.9 GB · 128K 27.7 GB
Q5_K_L
Seed OSS 36B Instruct
26.8 GB26.1 GB file
32K 30.8 GB · 128K 43.7 GB
Qwen3 30B A3B Thinking 2507
22.5 GB21.9 GB file
32K 24.1 GB · 128K 28.9 GB
BF16
Seed OSS 36B Instruct
72.9 GB72.3 GB file
32K 76.9 GB · 128K 89.8 GB
Qwen3 30B A3B Thinking 2507
61.5 GB61.1 GB file
32K 63.0 GB · 128K 67.8 GB
Seed OSS 36B Instruct
Ollama · Q4_K_M
ollama run hf.co/unsloth/Seed-OSS-36B-Instruct-GGUF:Q4_K_MQwen3 30B A3B Thinking 2507
Ollama · Q4_K_M
ollama run hf.co/unsloth/Qwen3-30B-A3B-Thinking-2507-GGUF:Q4_K_M| Spec | Seed OSS 36B Instruct | Qwen3 30B A3B Thinking 2507 |
|---|---|---|
| Publisher | ByteDance-Seed | Alibaba |
| Parameters | 36.2B | 30.5B |
| Active parameters | — | 3.4B per token |
| Architecture | SeedOssForCausalLM | Qwen3MoeForCausalLM |
| Context | 512K tokens | 256K tokens |
| Modalities | Chat, Function-calling | Chat, Function-calling |
| License | Apache 2.0 | Apache 2.0 |
| Released | Aug 2025 | Jul 2025 |
Frequently Asked Questions
Which is better, Seed OSS 36B Instruct or Qwen3 30B A3B Thinking 2507?
Seed OSS 36B Instruct and Qwen3 30B A3B Thinking 2507 match in llmrun Score, with values of 102.5 versus 100.6; both fit 24 GB at Q4_K_M; Qwen3 30B A3B Thinking 2507 is 6.0× faster on RTX 4090 (~173 vs ~29 tok/s). Choose Seed OSS 36B Instruct for ahead on the benchmarks and longer context window; choose Qwen3 30B A3B Thinking 2507 for fits a 16 GB GPU and 6.0× faster on RTX 4090.
Which needs less VRAM, Seed OSS 36B Instruct or Qwen3 30B A3B Thinking 2507?
At Q4_K_M, Seed OSS 36B Instruct requires 22.3 GB while Qwen3 30B A3B Thinking 2507 needs 18.7 GB, making Qwen3 30B A3B Thinking 2507 the easier model to run on your local setup.
Which is faster on RTX 4090?
Estimated decode speed on RTX 4090 is about 29 tok/s for Seed OSS 36B Instruct (Q4_K_M) and 173 tok/s for Qwen3 30B A3B Thinking 2507 (Q4_K_M), so Qwen3 30B A3B Thinking 2507 is faster. Estimates, ±20%.
Can I run Seed OSS 36B Instruct or Qwen3 30B A3B Thinking 2507 on a 16 GB GPU?
Seed OSS 36B Instruct fits at IQ2_M; Qwen3 30B A3B Thinking 2507 fits at Q3_K_M.
Which has a longer context window, Seed OSS 36B Instruct or Qwen3 30B A3B Thinking 2507?
Seed OSS 36B Instruct supports 524,288 tokens and Qwen3 30B A3B Thinking 2507 supports 262,144 tokens.