SAVRN
Search Contact SAVRN

SAVRN Model Hub · Comparisons

gemma-4-26B-A4B-it vs Qwen3.6-35B-A3B-NVFP4

Gemma-4-26B-A4B-it has 25.8B parameters and Qwen3.6-35B-A3B-NVFP4 has 24.6B parameters; both are released under Apache License 2.0; at 16-bit, gemma-4-26B-A4B-it needs about 61.9 GB (1x MI300X from $1.85 an hour) and Qwen3.6-35B-A3B-NVFP4 about 59.1 GB (1x MI300X from $1.85 an hour).

Published metadata for 2 models, each read from its own repository.
Field gemma-4-26B-A4B-it
google/gemma-4-26B-A4B-it
Qwen3.6-35B-A3B-NVFP4
unsloth/Qwen3.6-35B-A3B-NVFP4
Publisher Google Unsloth AI
Task Image and text to text Image and text to text
Modality Image and text Image and text
Parameters, as reported 25.8B parameters 24.6B parameters
Architecture Gemma4ForConditionalGeneration Qwen3_5MoeForConditionalGeneration
Library transformers transformers
Context length 262,144 tokens 262,144 tokens
Repository size 51.6 GB 26.5 GB
Artifact formats safetensors safetensors
License apache-2.0 apache-2.0
Access Open weights, no gate Open weights, no gate
Memory at 16-bit (weights and margin) 61.9 GB 59.1 GB
Cheapest GPUs at 16-bit, per hour 1x MI300X, $1.85 1x MI300X, $1.85
Memory at 4-bit (weights and margin) 15.5 GB 14.8 GB
Cheapest GPUs at 4-bit, per hour 1x MI300X, $1.85 1x MI300X, $1.85
Revision viewed 4d7ae4984b7d 739af1e7aac3
Downloads reported by the hub 9.8M 1.4M
Last observed 2026-09-18 2026-09-18

An evaluation row appears only where at least two of these models report the same benchmark with the same stated configuration, metric, unit and setup. Different evaluators stay named in each cell. Values are shown as reported: no unit conversion, no ranking.

Other Reported Results

These results are listed for each model on its own, because the conditions needed to compare them are not stated or do not match. Two results that leave a condition blank are not assumed to share it.

gemma-4-26B-A4B-it

BenchmarkConditionsResultReported byRevisionDate
Idavidrein/gpqa Task diamondMetric diamondComparison conditions not established 82.3 Model Card
Reported by a third party
Evaluated revision not stated 2026-04-02
MMMU/MMMU_Pro Task mmmu_pro_visionMetric mmmu_pro_visionComparison conditions not established 73.8 Model Card
Reported by a third party
Evaluated revision not stated 2026-05-12
TIGER-Lab/MMLU-Pro Task mmlu_proMetric mmlu_proComparison conditions not established 82.6 Model Card
Reported by a third party
Evaluated revision not stated 2026-04-02
cais/hle Task hleMetric hleSetup With searchComparison conditions not established 17.2 Model Card
Reported by a third party
Evaluated revision not stated 2026-04-02
llamaindex/ExtractBench Task longMetric longSetup Pipeline name: gemma4_26b_vllm_extract_oneshot_structured_output_file; self-hosted on vLLM, one-shot JSON-Schema structured outputComparison conditions not established 28.88 ExtractBench
Reported by a third party
Evaluated revision not stated 2026-08-26
llamaindex/ExtractBench Task meanMetric meanSetup Pipeline name: gemma4_26b_vllm_extract_oneshot_structured_output_file; self-hosted on vLLM, one-shot JSON-Schema structured outputComparison conditions not established 70.27 ExtractBench
Reported by a third party
Evaluated revision not stated 2026-08-26
llamaindex/ExtractBench Task mediumMetric mediumSetup Pipeline name: gemma4_26b_vllm_extract_oneshot_structured_output_file; self-hosted on vLLM, one-shot JSON-Schema structured outputComparison conditions not established 59.61 ExtractBench
Reported by a third party
Evaluated revision not stated 2026-08-26
llamaindex/ExtractBench Task shortMetric shortSetup Pipeline name: gemma4_26b_vllm_extract_oneshot_structured_output_file; self-hosted on vLLM, one-shot JSON-Schema structured outputComparison conditions not established 77.7 ExtractBench
Reported by a third party
Evaluated revision not stated 2026-08-26
llamaindex/ParseBench Task chartMetric chartSetup Pipeline name: gemma4_26b_vllm_with_layoutComparison conditions not established 14.2 ParseBench
Reported by a third party
Evaluated revision not stated 2026-04-14
llamaindex/ParseBench Task layoutMetric layoutSetup Pipeline name: gemma4_26b_vllm_with_layoutComparison conditions not established 59.2 ParseBench
Reported by a third party
Evaluated revision not stated 2026-04-14
llamaindex/ParseBench Task meanMetric meanSetup Pipeline name: gemma4_26b_vllm_with_layoutComparison conditions not established 58.5 ParseBench
Reported by a third party
Evaluated revision not stated 2026-04-14
llamaindex/ParseBench Task tableMetric tableSetup Pipeline name: gemma4_26b_vllm_with_layoutComparison conditions not established 70 ParseBench
Reported by a third party
Evaluated revision not stated 2026-04-14
llamaindex/ParseBench Task text_contentMetric text_contentSetup Pipeline name: gemma4_26b_vllm_with_layoutComparison conditions not established 83.8 ParseBench
Reported by a third party
Evaluated revision not stated 2026-04-14
llamaindex/ParseBench Task text_formattingMetric text_formattingSetup Pipeline name: gemma4_26b_vllm_with_layoutComparison conditions not established 65.1 ParseBench
Reported by a third party
Evaluated revision not stated 2026-04-14

SAVRN's Notes on Qwen3.6-35B-A3B-NVFP4

Of the 256 experts in each layer, 8 fire per token, so only a fraction of the 24.6B parameters touch any one token; memory is the constraint, since every expert has to be resident. The 4-bit row reads 12.3 GB of weights and 14.8 GB needed, and the publisher puts it on a 32 GB card. The cheapest listed setup, one MI300X with 192 GB at $1.85 an hour on-demand, leaves most of the card for a 262,144-token context and for image inputs.

The license is Apache 2.0, so commercial use and modification are permitted with notices retained. Before you commit, note that this is an NVFP4 quantization of Qwen/Qwen3.6-35B-A3B published by Unsloth AI, calibrated on the publisher's own dataset mixed with UltraChat, and that the publisher serves it through vLLM; no host prices for it appear in the SAVRN Index.

Questions

Which is larger, gemma-4-26B-A4B-it or Qwen3.6-35B-A3B-NVFP4?

gemma-4-26B-A4B-it (25.8B parameters) is larger than Qwen3.6-35B-A3B-NVFP4 (24.6B parameters), by the parameter counts their publishers report.

Which is cheaper to run, gemma-4-26B-A4B-it or Qwen3.6-35B-A3B-NVFP4?

At 4-bit, gemma-4-26B-A4B-it fits on 1x MI300X from $1.85 an hour and Qwen3.6-35B-A3B-NVFP4 on 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.

Can I use gemma-4-26B-A4B-it commercially?

Yes. gemma-4-26B-A4B-it is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.

Can I use Qwen3.6-35B-A3B-NVFP4 commercially?

Yes. Qwen3.6-35B-A3B-NVFP4 is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.

Related Comparisons