Knowledge
Directional only- Gemma 4 26B A4B
- 43.4
- Qwen3.5 397B
- 56.6
- Weighted basis
- 2 vs 4 rows
- Reading
- Directional only
Gemma 4 26B A4B vs Qwen3.5 397B
Updated August 18, 2026. Public scores include evidence status and uncertainty. They are not guarantees for a specific workload.
Decision reading
Qwen3.5 397B has the higher public score estimate, 57.46 versus 57.16, but the 90% score intervals overlap. Treat that as a lead, not a settled winner.
3 results are shared. Category rows based on different benchmark sets are marked directional and do not name a winner.
Which one for your work
Recommendations appear only when a shared evidence basis or an explicit operating constraint supports the call. Secondary and unsupported use cases stay disclosed below the initial list.
Long documents
Prompts that approach the documented context limit
Gemma 4 26B A4B
Gemma 4 26B A4B has the larger documented context window.
Confidence: documented
Show secondary and unsupported calls
Coding work
Code generation, repair, and software-engineering tasks
Not enough matched evidence
No shared weighted benchmark basis supports a winner.
Confidence: limited
Agentic work
Tool use, computer use, and multi-step task completion
Not enough matched evidence
No shared weighted benchmark basis supports a winner.
Confidence: limited
Chat turn cost
1K fresh input + 500 output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
Confidence: listed-rates
Cache-heavy agent loop cost
200K cached + 20K fresh input + 10K output tokens
Not enough matched evidence
The page does not recommend a cost winner because at least one model cannot fit the stated workload in one request. Qwen3.5 397B does not fit this workload in one request. Qwen3.5 397B has no published cached-input rate, so cached tokens use its listed input rate. Gemma 4 26B A4B has no comparable published API token rate.
Confidence: rate-fallback
Repository review cost
50K fresh input + 3K output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
Confidence: listed-rates
What is actually comparable
Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.
1 Gemma 4 26B A4B only
3 Shared
35 Qwen3.5 397B only
- Shared results
- 3
- Gemma 4 26B A4B only
- 1
- Qwen3.5 397B only
- 35
- Like-for-like categories
- 0 / 8
2 categories use different evidence sets. Those rows remain visible for coverage context but do not name a winner.
Category results, on a stated basis
Each row states whether both averages use the same weighted benchmark set. Directional and not-comparable rows remain visible, but they never receive a winner in this template.
Multimodal
Directional only- Gemma 4 26B A4B
- 73.8
- Qwen3.5 397B
- 79.6
- Weighted basis
- 1 vs 2 rows
- Reading
- Directional only
Agentic
Not comparable- Gemma 4 26B A4B
- Not measured
- Qwen3.5 397B
- 56.5
- Weighted basis
- 0 vs 2 rows
- Reading
- Not comparable
Coding
Not comparable- Gemma 4 26B A4B
- Not measured
- Qwen3.5 397B
- 66.5
- Weighted basis
- 0 vs 2 rows
- Reading
- Not comparable
Reasoning
Not comparable- Gemma 4 26B A4B
- Not measured
- Qwen3.5 397B
- 63.2
- Weighted basis
- 0 vs 1 rows
- Reading
- Not comparable
Math
Not comparable- Gemma 4 26B A4B
- Not measured
- Qwen3.5 397B
- 90.6
- Weighted basis
- 0 vs 2 rows
- Reading
- Not comparable
Multilingual
Not comparable- Gemma 4 26B A4B
- Not measured
- Qwen3.5 397B
- 84.7
- Weighted basis
- 0 vs 1 rows
- Reading
- Not comparable
Instruction following
Not comparable- Gemma 4 26B A4B
- Not measured
- Qwen3.5 397B
- 92.6
- Weighted basis
- 0 vs 1 rows
- Reading
- Not comparable
| Category | Gemma 4 26B A4B | Qwen3.5 397B | Weighted basis | Reading |
|---|---|---|---|---|
| Knowledge | 43.4 | 56.6 | Directional only2 vs 4 rows | Directional only |
| Multimodal | 73.8 | 79.6 | Directional only1 vs 2 rows | Directional only |
| Agentic | Not measured | 56.5 | Not comparable0 vs 2 rows | Not comparable |
| Coding | Not measured | 66.5 | Not comparable0 vs 2 rows | Not comparable |
| Reasoning | Not measured | 63.2 | Not comparable0 vs 1 rows | Not comparable |
| Math | Not measured | 90.6 | Not comparable0 vs 2 rows | Not comparable |
| Multilingual | Not measured | 84.7 | Not comparable0 vs 1 rows | Not comparable |
| Instruction following | Not measured | 92.6 | Not comparable0 vs 1 rows | Not comparable |
Shape of the matched evidence
Only shared public evidence is shown. Sparse evidence stays a ruled list rather than being closed into a radar shape.
Too few matched category axes support a radar. The ruled list below shows only shared benchmark results; positions use each benchmark’s normalized display scale when available.
HLE
Knowledge
Gemma 4 26B A4B: 17.2%Qwen3.5 397B: 28.7%Normalized gap 11.5Gemma 4 26B A4B source Qwen3.5 397B sourceMMLU-Pro
Knowledge
Gemma 4 26B A4B: 82.6%Qwen3.5 397B: 87.8%Normalized gap 5.2Gemma 4 26B A4B source Qwen3.5 397B sourceMMMU-Pro
Multimodal
What each workload costs
Three fixed token mixes turn per-token rates into comparable decisions. Each scenario states context fit and whether cached input had to fall back to the published list-input rate.
Chat turn
1K fresh input + 500 output tokens
- Gemma 4 26B A4B
- Self-hosted; infrastructure cost varies
- Fits in one request
- Qwen3.5 397B
- $0.0024
- Fits in one request
Gemma 4 26B A4B has no comparable published API token rate.
Repository review
50K fresh input + 3K output tokens
- Gemma 4 26B A4B
- Self-hosted; infrastructure cost varies
- Fits in one request
- Qwen3.5 397B
- $0.0408
- Fits in one request
Gemma 4 26B A4B has no comparable published API token rate.
Cache-heavy agent loop
200K cached + 20K fresh input + 10K output tokens
- Gemma 4 26B A4B
- Self-hosted; infrastructure cost varies
- Fits in one request
- Cached-input rate unavailable
- Qwen3.5 397B
- $0.168
- Does not fit in one request
- Cached input priced at the published list-input rate
Qwen3.5 397B does not fit this workload in one request. Qwen3.5 397B has no published cached-input rate, so cached tokens use its listed input rate. Gemma 4 26B A4B has no comparable published API token rate.
Specification differences
Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.
SpecificationGemma 4 26B A4BQwen3.5 397B
Context window
Maximum documented context; output-token limits may be lower.
Gemma 4 26B A4B
Qwen3.5 397B
128K
API model ID
Gemma 4 26B A4B
gemma-4-26b-a4b-it
Google Gemma Gemini API guideQwen3.5 397B
Not sourced
Cached-input rate
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
Gemma 4 26B A4B
No comparable hosted API rate
Qwen3.5 397B
Not published
Documented inputs
Gemma 4 26B A4B
text, image
Google Gemma 4 model documentationQwen3.5 397B
Not sourced
Documented outputs
Gemma 4 26B A4B
Qwen3.5 397B
Not sourced
Provider availability
Gemma 4 26B A4B
Generally Available · Gemini API, Google AI Studio, open weights
Google Gemma Gemini API guideQwen3.5 397B
Not sourced
Reasoning profile
Gemma 4 26B A4B
Reasoning
Qwen3.5 397B
Non-Reasoning
Weight access
Gemma 4 26B A4B
Open Weight
Qwen3.5 397B
Open Weight
License
Gemma 4 26B A4B
Open Weight
Qwen3.5 397B
Open Weight
Release date
Gemma 4 26B A4B
2026-04-02
Qwen3.5 397B
2026-02-16
If you already use one of these models
- Deployment change
- The models list different providers, so authentication, endpoint behavior, limits, and feature support may change.
- Quality signal
- Qwen3.5 397B has the higher public score estimate, 57.46 versus 57.16, but the 90% score intervals overlap.
- Workload cost
- A complete comparable API-rate estimate is not available for both models.
- Context tradeoff
- Gemma 4 26B A4B has the larger documented window (256K).
Run the same representative tasks against both endpoints before changing production traffic.
Benchmark evidence
The full public result ledger is available for audit without forcing a wide desktop table onto a phone.
Browse raw public benchmark evidence39 rows
Agentic
Terminal-Bench 2.0
Gemma 4 26B A4B—Qwen3.5 397B52.5%SourceNot directly comparable
BrowseComp
Gemma 4 26B A4B—Qwen3.5 397B62%SourceNot directly comparable
Claw-Eval
Gemma 4 26B A4B—Qwen3.5 397B56.8%SourceNot directly comparable
QwenClawBench
Gemma 4 26B A4B—Qwen3.5 397B51.8%SourceNot directly comparable
τ³-bench results
Gemma 4 26B A4B—Qwen3.5 397B68.4%SourceNot directly comparable
VITA-Bench
Gemma 4 26B A4B—Qwen3.5 397B43.7%SourceNot directly comparable
DeepPlanning
Gemma 4 26B A4B—Qwen3.5 397B37.6%SourceNot directly comparable
Toolathlon
Gemma 4 26B A4B—Qwen3.5 397B36.3%SourceNot directly comparable
MCP Atlas
Gemma 4 26B A4B—Qwen3.5 397B46.1%SourceNot directly comparable
MCP-Tasks
Gemma 4 26B A4B—Qwen3.5 397B74.2%SourceNot directly comparable
WideResearch
Gemma 4 26B A4B—Qwen3.5 397B74.0%SourceNot directly comparable
Gert Labs
Gemma 4 26B A4B—Qwen3.5 397B46.76%SourceNot directly comparable
ResearchClawBench
Gemma 4 26B A4B—Qwen3.5 397B14.2%SourceNot directly comparable
Coding
Reasoning
Knowledge
MMLU-Pro
Gemma 4 26B A4B82.6%SourceQwen3.5 397B87.8%SourceQwen3.5 397B leads this result
HLE
Gemma 4 26B A4B17.2%SourceQwen3.5 397B28.7%SourceQwen3.5 397B leads this result
HLE w/o tools
Gemma 4 26B A4B8.7%SourceQwen3.5 397B—Not directly comparable
GPQA
Gemma 4 26B A4B—Qwen3.5 397B88.4%SourceNot directly comparable
SuperGPQA
Gemma 4 26B A4B—Qwen3.5 397B70.4%SourceNot directly comparable
MMLU-Redux
Gemma 4 26B A4B—Qwen3.5 397B94.9%SourceNot directly comparable
C-Eval
Gemma 4 26B A4B—Qwen3.5 397B93%SourceNot directly comparable
Math
AIME26
Gemma 4 26B A4B—Qwen3.5 397B93.3%SourceNot directly comparable
HMMT Feb 2025
Gemma 4 26B A4B—Qwen3.5 397B94.8%SourceNot directly comparable
HMMT Nov 2025
Gemma 4 26B A4B—Qwen3.5 397B92.7%SourceNot directly comparable
HMMT Feb 2026
Gemma 4 26B A4B—Qwen3.5 397B87.9%SourceNot directly comparable
MMAnswerBench
Gemma 4 26B A4B—Qwen3.5 397B80.9%SourceNot directly comparable
Multilingual
Multimodal
MMMU-Pro
Gemma 4 26B A4B73.8%SourceQwen3.5 397B79%SourceQwen3.5 397B leads this result
MathVision
Gemma 4 26B A4B—Qwen3.5 397B88.6%SourceNot directly comparable
CharXiv
Gemma 4 26B A4B—Qwen3.5 397B80.8%SourceNot directly comparable
VideoMMMU
Gemma 4 26B A4B—Qwen3.5 397B84.7%SourceNot directly comparable
ScreenSpot Pro
Gemma 4 26B A4B—Qwen3.5 397B65.6%SourceNot directly comparable
V*
Gemma 4 26B A4B—Qwen3.5 397B95.8%SourceNot directly comparable
Instruction following
IFEval
Gemma 4 26B A4B—Qwen3.5 397B92.6%SourceNot directly comparable
Frequently asked questions
Which is better, Gemma 4 26B A4B or Qwen3.5 397B?
Qwen3.5 397B has the higher public score estimate, 57.46 versus 57.16, but the 90% score intervals overlap. The higher estimate is not a decisive winner because the uncertainty ranges overlap.
Which is better for coding, Gemma 4 26B A4B or Qwen3.5 397B?
The published evidence does not provide a shared weighted coding basis for both models, so BenchLM does not name a coding winner.
Which is better for agentic tasks, Gemma 4 26B A4B or Qwen3.5 397B?
The published evidence does not provide a shared weighted agentic tasks basis for both models, so BenchLM does not name a agentic tasks winner.
Which costs less, Gemma 4 26B A4B or Qwen3.5 397B?
Both models do not have comparable published API token rates, so this page does not name a universal price winner.
Which has the larger context window, Gemma 4 26B A4B or Qwen3.5 397B?
Gemma 4 26B A4B has the larger documented context window: 256K, compared with 128K.
Related comparisons
Last updated August 18, 2026
Watch Gemma 4 26B A4B vs Qwen3.5 397B
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 2,000+ readers.