DiffusionGemma 26B A4B
DiffusionGemma 26B A4BAlternatives?
CheaperGPT OSS 120BhighOpen+81%-20%$0.092Context: 256K → 131K+2BetterGPT-6 Lunalow+87%+167%$0.005Function calling only with none · Chat Completions+3
OpenQwen3.8 27Bxhigh×3.1+1,175%$1.01Memory (4-bit): 15 GB → 17 GB+5CheaperGPT OSS 20BhighOpensame-52%$0.011Context: 256K → 131K+1
CheaperNemotron 3.5 LightningthinkingOpensame-35%$0.079Memory (4-bit): 15 GB → 20 GB+1GPT OSS 120BhighOpen+81%-20%$0.092Context: 256K → 131K+2GPT OSS 20BhighOpensame-52%$0.011Context: 256K → 131K+1Nemotron 3.5 LightningthinkingOpensame-35%$0.079Memory (4-bit): 15 GB → 20 GB+1Qwen3.5 4BthinkingOpensame-20%$0.03 / $0.15
Granite 4.2 3BthinkingOpen-39%-30%$0.006Context: 256K → 128K+1GPT-6 Lunalow+87%+167%$0.005Function calling only with none · Chat Completions+3GPT-5.6 Lunalow+83%+500%$0.01
MiMo v2.5thinkingOpen×2.5+133%$0.019Memory (4-bit): 15 GB → 186 GB
Claude Haiku 5.5low×2.5+167%$0.024Remove sampling parameters+5
Mercury 2high+58%+400%$0.039Context: 256K → 128K+1Qwen3.8 27BxhighOpen×3.1+1,175%$1.01Memory (4-bit): 15 GB → 17 GB+5Qwen 3.5 27BthinkingOpen×2.2+1,000%$0.30 / $2.40Memory (4-bit): 15 GB → 17 GBGemma 4 31BthinkingOpen×2.2+113%$0.10 / $0.34Memory (4-bit): 15 GB → 19 GBIK2 Horizon MoVA 36B A4BthinkingOpen×2.2+2,233%$1 / $4Memory (4-bit): 15 GB → 22 GB+1Qwen3.6 27BthinkingOpen×2.1+1,700%$0.62Memory (4-bit): 15 GB → 17 GBGemini 4 Argonhigh×5.3—*$1.99Synchronous output stops at 65,536+2Gemini 3.1 Pro Previewthinking×4.3—$1.30Cannot use both thinking controlsGemini 3.8 Flashmedium×3.9—*$0.93Sampling parameters are ignored+1Gemini 3.6 Flashhigh×3.8—*$1.60Gemma 4 31BthinkingOpen×2.2—$0.10 / $0.34Memory (4-bit): 15 GB → 19 GBGemini 2.5 Prothinking×2.1—$0.33Gemini 2.5 Pro Preview (May' 25)thinking+84%——No image inputGemma 4 26B A4BthinkingOpen+78%—$0.15 / $0.60All Google models →Run it yourself
estimateBenchmarks
in Chat & KnowledgeFAQ
How much does DiffusionGemma 26B A4B cost?›
$0.05 per 1M input tokens and $0.15 per 1M output tokens (cheapest host: NanoGPT).
Is DiffusionGemma 26B A4B free to use?›
There is no free API; the cheapest paid price is $0.05 per 1M input tokens. The weights are open, so you can run it yourself.
How good is DiffusionGemma 26B A4B, and what is it best at?›
#179 of 334 in Chat & Knowledge, #122 of 156 in Coding, #122 of 175 in Agents & Tool Use, #86 of 245 in Writing & Language, #78 of 124 in Vision & Documents.
Is DiffusionGemma 26B A4B better than Nanbeige4.1 3B?›
On Humanity's Last Exam, DiffusionGemma 26B A4B scores 10.8 vs 10.9 — about the same.
How can I use DiffusionGemma 26B A4B?›
Through the Google API.
What is the context window of DiffusionGemma 26B A4B?›
256K tokens.
When was DiffusionGemma 26B A4B released? Is there a newer version?›
Released 2026-06-10. It is the current version in its line.
Is DiffusionGemma 26B A4B open source?›
Yes, the weights are open (Apache-2.0 licence).
Can I run DiffusionGemma 26B A4B locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can DiffusionGemma 26B A4B read images?›
It accepts text and images as input.
What is DiffusionGemma 26B A4B?›
DiffusionGemma 26B A4B is an open-weights language model from Google, released Jun 2026.
What are the best alternatives to DiffusionGemma 26B A4B?›
GPT OSS 120B (cheaper), GPT-6 Luna (better), Qwen3.8 27B (open weights).