DeepSeek R1 0528 Qwen3 8B
May 2025
DeepSeek R1 0528 Qwen3 8BCompare inon
Benchmarks of DeepSeek R1 0528 Qwen3 8B15
BenchmarkByScoreRank
Overall1
Knowledge4
Science2
Coding1
Documents1
Legacy · saturated or no longer reported6
No benchmark matches
Rankingin thinking mode?
Alternatives?
ModelScoreCostPer taskWhat changes
CheaperQwen3.5 4BthinkingOpen+62%-11%$0.03 / $0.15
BetterGPT-6 Lunalow×2.7+196%$0.005Function calling only with none · Chat Completions+3IOpenK2 Horizon 7Bthinking×2.5+678%$0.30 / $1.20Memory (4-bit): 4.1 GB → 4.5 GB
CheaperNemotron 3.5 LightningthinkingOpen+59%-28%$0.079Memory (4-bit): 4.1 GB → 20 GBCheaperGPT OSS 120BhighOpen+44%-11%$0.092Memory (4-bit): 4.1 GB → 70 GBModelScoreCostPer taskWhat changes
Qwen3.5 4BthinkingOpen+62%-11%$0.03 / $0.15Nemotron 3.5 LightningthinkingOpen+59%-28%$0.079Memory (4-bit): 4.1 GB → 20 GBGPT OSS 120BhighOpen+44%-11%$0.092Memory (4-bit): 4.1 GB → 70 GBGPT OSS 20BlowOpen+23%-47%$0.018 / $0.09Memory (4-bit): 4.1 GB → 13 GB
Granite 4.2 3BthinkingOpensame-22%$0.006ModelScoreCostPer taskWhat changes
GPT-6 Lunalow×2.7+196%$0.005Function calling only with none · Chat Completions+3GPT-5.6 Lunalow×2.6+567%$0.01
Mistral Small 4thinkingOpen+39%+289%$0.015Memory (4-bit): 4.1 GB → 72 GB
MiMo v2.5thinkingOpen×2.7+159%$0.019Memory (4-bit): 4.1 GB → 186 GBGranite 4.2 8BthinkingOpen+37%+59%$0.024Memory (4-bit): 4.1 GB → 4.4 GBModelScoreCostPer taskWhat changes
IK2 Horizon 7BthinkingOpen×2.5+678%$0.30 / $1.20Memory (4-bit): 4.1 GB → 4.5 GBIK2 Horizon 3.7BthinkingOpen+93%+678%$0.30 / $1.20
Gemma 4 12BthinkingOpen+76%+122%$0.10 / $0.30Memory (4-bit): 4.1 GB → 6 GBSApriel v1.5 15B ThinkerOpen+71%——Memory (4-bit): 4.1 GB → 7.4 GBSApriel v1.6 15B ThinkerOpen+66%——Memory (4-bit): 4.1 GB → 7.4 GBModelScoreCostPer taskWhat changes
DeepSeek V4.1 Flash 1017maxOpen×5.6—$0.26Memory (4-bit): 4.1 GB → 458 GB+5DeepSeek V4 Pro 0813non-thinkingOpen×2.5—$0.36Memory (4-bit): 4.1 GB → 990 GB+5DeepSeek V3.2 SpecialethinkingOpen+79%——Memory (4-bit): 4.1 GB → 411 GBDeepSeek V3.1 Terminusnon-thinkingOpen+72%—$0.27 / $0.95Memory (4-bit): 4.1 GB → 411 GBDeepSeek V3.1non-thinkingOpen+70%—$0.25 / $0.95Memory (4-bit): 4.1 GB → 411 GBR1 0528thinkingOpen+62%—$0.39 / $1.63Memory (4-bit): 4.1 GB → 411 GBDeepSeek V3 0324non-thinkingOpen+20%—$0.021Memory (4-bit): 4.1 GB → 411 GBDeepSeek V3non-thinkingOpensame—$0.02Memory (4-bit): 4.1 GB → 411 GBAll DeepSeek models →Run it yourself
estimateFits on✓ 8 GB (phone, laptop) · compressed✓ 16 GB · 8-bit✓ 24 GB GPU · full quality✓ 48 GB GPU · full quality✓ 80 GB GPU · full quality
LicenceMITfree for commercial use
Benchmarks
all categoriesChat & Knowledge
Humanity's Last Exam5.9#252
AA Omniscience Index-63.6#277
CritPt (physics)0#180
AA Intelligence Index 4.3.2≈8.1#246
AA Omniscience · Accuracy12.2#267
GPQA Diamond61.2#225
AIME 2025 · AA archived63.7#77
Coding
LiveCodeBench · AA archived51.3#89
Agents & Tool Use
τ²-bench · AA0#232
Writing & Language
IFBench (instructions)19.9#244
FAQ
How much does DeepSeek R1 0528 Qwen3 8B cost?›
$0.06 per 1M input tokens and $0.09 per 1M output tokens (cheapest host: Novita).
Is DeepSeek R1 0528 Qwen3 8B free to use?›
There is no free API; the cheapest paid price is $0.06 per 1M input tokens. The weights are open, so you can run it yourself.
How good is DeepSeek R1 0528 Qwen3 8B, and what is it best at?›
#250 of 334 in Chat & Knowledge, #195 of 239 in Coding, #229 of 242 in Agents & Tool Use, #241 of 245 in Writing & Language.
Is DeepSeek R1 0528 Qwen3 8B better than Qwen3 VL 8B (Reasoning)?›
On AA Intelligence Index 4.3.2, DeepSeek R1 0528 Qwen3 8B scores 8.1 vs 8.2 — about the same.
How can I use DeepSeek R1 0528 Qwen3 8B?›
Through the API of Novita.
What is the context window of DeepSeek R1 0528 Qwen3 8B?›
33K tokens.
When was DeepSeek R1 0528 Qwen3 8B released? Is there a newer version?›
Released 2025-05-29. It has been superseded by DeepSeek V3 0324.
Is DeepSeek R1 0528 Qwen3 8B open source?›
Yes, the weights are open (MIT licence).
Can I run DeepSeek R1 0528 Qwen3 8B locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can DeepSeek R1 0528 Qwen3 8B read images?›
No, it accepts text only.
What is DeepSeek R1 0528 Qwen3 8B?›
DeepSeek R1 0528 Qwen3 8B is an open-weights language model from DeepSeek, released May 2025.
What are the best alternatives to DeepSeek R1 0528 Qwen3 8B?›
Qwen3.5 4B (cheaper), GPT-6 Luna (better), K2 Horizon 7B (open weights).