Qwen
·Released on Feb 25, 2026Qwen3.5-35B-A3B API Benchmarks, Pricing & Provider Data
Compare Qwen3.5-35B-A3B with another model
Choose a model to open its comparison page.
LLM
Qwen3.5-35B-A3B benchmark, API pricing, and provider data cover 30 API providers, with prices starting at $0.022/M.
The Qwen3.5 Series 35B-A3B is a native vision-language model designed with a hybrid architecture that integrates linear attention mechanisms and a sparse mixture-of-experts model, achieving higher inf...
Quality
#78of 100
48.0
LMSpeed score
Cost
#44of 180
$0.022/ 1M · 8:1 in:out
$0.0035 in · $0.018 out
Category Performance
Observed-capability estimates with uncertainty reported separately from independent benchmark families.
- Coverage
- 7 / 8
- Methodology
- V3.0
#1Instruction following51.5Estimated2/4 Measured dimensions
80% interval: 39.5–63.4
constraint following49.6
ifeval · ifeval
structured outputPrior only
novel instruction generalization53.4
ifbench · aa_if_bench
long multiturn instructionPrior only
#2Reasoning51.2RatedGlobal rank #363/4 Measured dimensions
80% interval: 42.9–59.5
abstract logicPrior only
scientific causal52.5
critpt · critpt / gpqa · gpqa / hle · hle
multistep constraints56.5
mmlu_pro · mmlu_pro
evidence verification44.8
lcr · lcr / longbench · long_bench_v2
#3Multimodal45.8Estimated2/4 Measured dimensions
80% interval: 33.8–57.9
perception ocr49.9
v_star · v_star
document spatialPrior only
visual reasoning41.8
mathvision · math_vision
video actionPrior only
#4Agents42.8RatedGlobal rank #513/4 Measured dimensions
80% interval: 34.1–51.5
planning39.1
deep_planning · gert_labs
tool use52.1
tau · tau2_bench
environment execution37.2
browsecomp · browse_comp / osworld · os_world_verified / terminalbench · benchlm_agentic_terminal_bench2
recovery reliabilityPrior only
#5Coding42Estimated2/4 Measured dimensions
80% interval: 30.8–53.2
code generation50.8
scicode · scicode
repository engineeringPrior only
debugging testing33.2
swe_rebench · swe_rebench / swe_verified · swe_verified
tooling qualityPrior only
#6Multilingual41.6Provisional1/4 Measured dimensions
80% interval: 25.1–58.0
cross language understandingPrior only
multilingual generationPrior only
reasoning transfer41.6
mmlu_prox · mmlu_pro_x
low resource robustnessPrior only
#7Knowledge38Provisional1/4 Measured dimensions
80% interval: 21.8–54.3
broad knowledgePrior only
professional knowledge38
supergpqa · super_gpqa
factualityPrior only
retrieval open bookPrior only
No data:Math
Specifications
Input and output token limits for this model, plus how it ranks on long-context understanding.
INPUT
262.1Ktokens
≈ 314.6 pages of text
OUTPUT
262.1Ktokens
8K128K1M4M
262.1K
Features
Technical Details
- Input
- Output
- Total parameters
- 35B
- Active parameters
- 3B
- Released
- Feb 2026
- Tokenizer
- Qwen3
- Architecture
- text+image+video->text
- Moderated
- No
- Supported parameters
- frequency_penaltyinclude_reasoninglogit_biaslogprobsmax_tokensmin_ppresence_penaltyreasoningrepetition_penaltyresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_ktop_logprobstop_p
Rankings
Excels at
It's decent at
Falls behind in
Detailed scores
Updated: Aug 23, 2026Overall
undefined metric} other undefined metrics}}
Overall score48.0#78 / 100
Overall
undefined metric} other undefined metrics}}
Benchmark
LMSpeed rank
Score
Source
Updated
Confidence
Speed & latency
undefined metric} other undefined metrics}}
Output speed169.5 tok/s#13 / 74Time to first token1.42 s#41 / 74
Speed & latency
undefined metric} other undefined metrics}}
Benchmark
LMSpeed rank
Score
Source
Updated
Confidence
Output speed
LMSpeed rank#13
Score169.5 tok/s
Sourceartificialanalysis.ai
UpdatedAug 23, 2026
Confidence4
Time to first token
LMSpeed rank#41
Score1.42 s
Sourceartificialanalysis.ai
UpdatedAug 23, 2026
Confidence4
Pricing
undefined metric} other undefined metrics}}
Input price$0.250/M#44 / 180Output price$2.00/M#70 / 180
Pricing
undefined metric} other undefined metrics}}
Benchmark
LMSpeed rank
Score
Source
Updated
Confidence
Agents
V3.0undefined metric} other undefined metrics}} · Rated
Score42.8#5180% interval34.1–51.53/4 Measured dimensionsAgentic score12.9#68 / 69Terminal-Bench 2.040.5#52 / 53
Agents
V3.0undefined metric} other undefined metrics}} · Rated
Dimensions and evidence
Planning & decomposition39.1
undefined family} other undefined families}} · undefined metric} other undefined metrics}}
deep_planning · gert_labs · z -1.44 · q 1.00
Tool use52.1
undefined family} other undefined families}} · undefined metric} other undefined metrics}}
tau · tau2_bench · z 0.16 · q 1.00
Environment & long-horizon execution37.2
undefined family} other undefined families}} · undefined metric} other undefined metrics}}
browsecomp · browse_comp · z -1.33 · q 1.00
osworld · os_world_verified · z -1.49 · q 1.00
terminalbench · benchlm_agentic_terminal_bench2 · z -1.40 · q 1.00
Recovery & completion reliabilityPrior only
Benchmark
LMSpeed rank
Score
Source
Updated
Confidence
Coding
V3.0undefined metric} other undefined metrics}} · Estimated
Score4280% interval30.8–53.22/4 Measured dimensionsSciCode29.3%#158 / 204Coding score47.0#63 / 81
Coding
V3.0undefined metric} other undefined metrics}} · Estimated
Dimensions and evidence
Code generation50.8
undefined family} other undefined families}} · undefined metric} other undefined metrics}}
scicode · scicode · z -0.04 · q 1.00
Repository engineeringPrior only
Debugging & testing33.2
undefined family} other undefined families}} · undefined metric} other undefined metrics}}
swe_rebench · swe_rebench · z -1.98 · q 1.00
swe_verified · swe_verified · z -1.36 · q 1.00
Tool-assisted development & qualityPrior only
Benchmark
LMSpeed rank
Score
Source
Updated
Confidence
SciCodeV3 evidence
LMSpeed rank#158
Score29.3%
Sourceartificialanalysis.ai
UpdatedAug 23, 2026
Confidence4
Reasoning
V3.0undefined metric} other undefined metrics}} · Rated
Score51.2#3680% interval42.9–59.53/4 Measured dimensionsMMLU-Pro85.3%#27 / 129GPQA81.9%#85 / 211
Reasoning
V3.0undefined metric} other undefined metrics}} · Rated
Dimensions and evidence
Abstract logicPrior only
Scientific & causal reasoning52.5
undefined family} other undefined families}} · undefined metric} other undefined metrics}}
critpt · critpt · z -0.47 · q 1.00
gpqa · gpqa · z 0.46 · q 1.00
hle · hle · z 0.33 · q 1.00
Multi-step constraints56.5
undefined family} other undefined families}} · undefined metric} other undefined metrics}}
mmlu_pro · mmlu_pro · z 0.67 · q 1.00
Evidence integration & verification44.8
undefined family} other undefined families}} · undefined metric} other undefined metrics}}
lcr · lcr · z -0.11 · q 1.00
longbench · long_bench_v2 · z -1.02 · q 1.00
Benchmark
LMSpeed rank
Score
Source
Updated
Confidence
Knowledge
V3.0undefined metric} other undefined metrics}} · Provisional
Score3880% interval21.8–54.31/4 Measured dimensionsKnowledge score75.6#23 / 69SuperGPQA63.4#15 / 16
Knowledge
V3.0undefined metric} other undefined metrics}} · Provisional
Dimensions and evidence
Broad knowledgePrior only
Professional knowledge38
undefined family} other undefined families}} · undefined metric} other undefined metrics}}
supergpqa · super_gpqa · z -1.34 · q 1.00
FactualityPrior only
Retrieval & open-book usePrior only
Benchmark
LMSpeed rank
Score
Source
Updated
Confidence
Artificial Analysis Intelligence Index
LMSpeed rank#72
Score29.9
Sourcebenchlm.ai
UpdatedAug 20, 2026
Confidence3
AA-Omniscience Hallucination Rate
LMSpeed rank#30
Score85.4
Sourcebenchlm.ai
UpdatedAug 20, 2026
Confidence3
Math
V3.0undefined metric} other undefined metrics}} · No data
No data80% interval30.8–69.20/4 Measured dimensions
Math
V3.0undefined metric} other undefined metrics}} · No data
Dimensions and evidence
Foundational mathPrior only
Competition mathPrior only
Advanced proofsPrior only
Applied & tool-assisted mathPrior only
Multilingual
V3.0undefined metric} other undefined metrics}} · Provisional
Score41.680% interval25.1–58.01/4 Measured dimensionsMultilingual score21.1#10 / 11MMLU-ProX81.0#10 / 11
Multilingual
V3.0undefined metric} other undefined metrics}} · Provisional
Dimensions and evidence
Cross-language understandingPrior only
Multilingual generationPrior only
Reasoning transfer41.6
undefined family} other undefined families}} · undefined metric} other undefined metrics}}
mmlu_prox · mmlu_pro_x · z -0.92 · q 1.00
Low-resource robustnessPrior only
Benchmark
LMSpeed rank
Score
Source
Updated
Confidence
Multimodal
V3.0undefined metric} other undefined metrics}} · Estimated
Score45.880% interval33.8–57.92/4 Measured dimensionsMMMU81.4#6 / 7MMVU72.3#5 / 5
Multimodal
V3.0undefined metric} other undefined metrics}} · Estimated
Dimensions and evidence
Perception & OCR49.9
undefined family} other undefined families}} · undefined metric} other undefined metrics}}
v_star · v_star · z -0.27 · q 1.00
Document & spatial understandingPrior only
Visual reasoning41.8
undefined family} other undefined families}} · undefined metric} other undefined metrics}}
mathvision · math_vision · z -1.01 · q 1.00
Video & grounded actionPrior only
Benchmark
LMSpeed rank
Score
Source
Updated
Confidence
Instruction following
V3.0undefined metric} other undefined metrics}} · Estimated
Score51.580% interval39.5–63.42/4 Measured dimensionsInstruction Following score85.2#16 / 25IFEval91.9#11 / 16
Instruction following
V3.0undefined metric} other undefined metrics}} · Estimated
Dimensions and evidence
Constraint following49.6
undefined family} other undefined families}} · undefined metric} other undefined metrics}}
ifeval · ifeval · z -0.26 · q 1.00
Structured outputPrior only
Novel-instruction generalization53.4
undefined family} other undefined families}} · undefined metric} other undefined metrics}}
ifbench · aa_if_bench · z 0.17 · q 1.00
Multi-turn & long instructionsPrior only
Benchmark
LMSpeed rank
Score
Source
Updated
Confidence
OpenRouter endpoints
7 endpointsThird-party OpenRouter endpoint data, shown separately from LMSpeed measurements. Some 30-minute live performance fields only appear after syncing with an OpenRouter API key.
| Provider endpoint | Input | Output | 1d uptime | 30m latency | 30m throughput | Context / output |
|---|---|---|---|---|---|---|
Parasail parasail/fp8 | $0.150/M | $1/M | 100.0% | — | — | undefined tokens / undefined tokens |
AtlasCloud atlas-cloud/fp8 | $0.225/M | $1.80/M | 99.9% | — | — | undefined tokens / undefined tokens |
Venice venice | $0.313/M | $1.25/M | 99.9% | — | — | undefined tokens / undefined tokens |
CoreWeave coreweave/fp8 | $0.250/M | $1.25/M | 99.9% | — | — | undefined tokens / undefined tokens |
DeepInfra deepinfra/fp8 | $0.140/M | $1/M | 99.6% | — | — | undefined tokens / undefined tokens |
Alibaba alibaba | $0.163/M | $1.30/M | 99.1% | — | — | undefined tokens / undefined tokens |
SiliconFlow siliconflow/fp8 | $0.240/M | $1.80/M | 97.4% | — | — | undefined tokens / undefined tokens |
Pricing Comparison
Compare Qwen3.5-35B-A3B API pricing across 30 providers. Prices range from $0.022/M to $0.510/M. Zhongzhuan Chat offers the lowest rate at $0.022/M.
| Provider | Health | Model Variant | Group | Input ($/M) | Output ($/M) | Speed (t/s) | First token | Audit |
|---|---|---|---|---|---|---|---|---|
L1 99% | qwen3.5-35b-a3b | default | -80%$0.051/M | -80%$0.408/M | — | — | — | |
L1 99% | qwen3.5-35b-a3b | default | -80%$0.051/M | -80%$0.408/M | — | — | — | |
L1 100% | qwen3.5-35b-a3b | default | -89%$0.027/M | -89%$0.219/M | — | — | — |
Alternatives & Similar Models
GPT-5.4
gpt-5-4
OpenAI GPT-5.4 extends the GPT-5 family with stronger instruction following, deeper tool use, and improved performance on coding, math, and long-document analysis.
3 shared providers
GPT-5.3 Codex
gpt-5-3-codex
OpenAI GPT-5.3 Codex is a code-specialized variant in the GPT-5 series, optimized for code generation, debugging, and software development tasks.
3 shared providers
GPT-5.2
gpt-5-2
OpenAI GPT-5.2 is a GPT-5 series model emphasizing advanced reasoning, multimodal understanding, and high-quality outputs for complex enterprise workloads.
3 shared providers
GPT-5.4 Mini
gpt-5-4-mini
OpenAI GPT-5.4 Mini is a compact language model in the GPT-5 series, optimized for quick responses and high throughput.
3 shared providers
Claude Opus 4.6
claude-opus-4-6
Anthropic Claude Opus 4.6 is the most capable Claude Opus tier, optimized for complex analysis, long-horizon coding, and high-stakes enterprise reasoning workloads.
3 shared providers
Claude Sonnet 4.6
claude-sonnet-4-6
Anthropic Claude Sonnet 4.6 extends the Sonnet line with improved tool use, coding reliability, and long-context performance for everyday production workloads.
3 shared providers
Frequently Asked Questions
- What benchmark data does Qwen3.5-35B-A3B include?
- LMSpeed shows Qwen3.5-35B-A3B benchmark context, API price, output speed, first-token latency, and provider data across 30 providers when those signals are available.
- What is the Qwen3.5-35B-A3B API price?
- Qwen3.5-35B-A3B has pricing from 30 providers, ranging from $0.022/M to $0.510/M. Zhongzhuan Chat has the lowest listed price.
- What does the Qwen3.5-35B-A3B API pricing table include?
- The Qwen3.5-35B-A3B API pricing table compares 30 providers by input price, output price, free tier status, speed, first-token latency, and recent health data when available.
- Which provider has the cheapest Qwen3.5-35B-A3B API pricing?
- Zhongzhuan Chat currently has the lowest listed Qwen3.5-35B-A3B price at $0.022/M across 30 providers.
- Is Qwen3.5-35B-A3B API free?
- Qwen3.5-35B-A3B does not currently have a free API tier on LMSpeed. All 30 providers charge per token.