Cheaper alternatives to Qwen3.5-35B-A3B

Qwen LMArena 1395.6 $0.500/M on a balanced workload prices as of 2026-08-28

1 model is both better and cheaper, giving up nothing.

Of the 133 models carrying both an independent score and a published price at the same delivery mode, 22 score at least as high as Qwen3.5-35B-A3B and cost no more under at least one workload. 1 matches or beats it on every capability we hold — context, maximum output, input modes, tool use and extended reasoning. 21 do not, and every row names what it drops.

Dominance is not a property of a model. It is a property of a model, a capability lens and a workload, and all three are named on every row. Scores are LMArena Elo, used under CC BY 4.0 from the official dataset; prices are blended per million tokens. A higher score is not a drop-in replacement.

This page is built from that comparison and nothing else. When no model is both better and cheaper than Qwen3.5-35B-A3B, it is not generated.

Where it holds

Workload	Qwen3.5-35B-A3B $/M	Better and cheaper	With a trade
Balanced	$0.500	0	13
Summarise	$0.300	1	18
Chat	$0.650	1	17
Code gen	$0.850	1	15
Agentic	$0.400	1	21

Workload mixes are defined on the methodology page. Cached input is priced at the cached rate, and reasoning tokens at the worse of the reasoning and output rates.

What you would be replacing

Intelligence	1395.6
Context window	262K
Max output	236K
Input modes	text, image, video
Tool use	yes
Extended reasoning	yes

A replacement has to clear every line above, not just the score. Full record for Qwen3.5-35B-A3B.

Better and cheaper, nothing given up 1

Each of these matches or beats Qwen3.5-35B-A3B on context, maximum output, input modes, tool use and reasoning, scores at least as high, and costs no more.

#	Model	Intelligence	$/M	Holds under
1	MiniMax M3 MiniMax	1434.8 +39.2	$0.277 summarise	Summarise −8% Chat −10% Code gen −3% Agentic −27%

Cheaper and higher-scoring, but you give something up 21

These score at least as high and cost no more on the two plotted axes, and lose something that is not on them. Read the last column before switching.

Model	Intelligence	$/M	Holds under	What you give up
Gemma 4 31B Google	1441.7 +46.1	$0.152	Balanced −70% Summarise −70% Chat −73% Code gen −72% Agentic −74%	236K → 16K max output
Hy3 Tencent	1441.2 +45.6	$0.144	Balanced −71% Summarise −74% Chat −75% Code gen −73% Agentic −79%	236K → 128K max output no image, video input
Gemma 4 26B A4B Google	1434.6 +39	$0.138	Balanced −73% Summarise −72% Chat −73% Code gen −73% Agentic −72%	236K → 16K max output
DeepSeek V4 Flash 0423 DeepSeek	1431.6 +36	$0.101	Balanced −80% Summarise −78% Chat −86% Code gen −85% Agentic −86%	no image, video input
MiMo-V2.5 Xiaomi	1427.3 +31.7	$0.175	Balanced −65% Summarise −64% Chat −76% Code gen −75% Agentic −80%	236K → 131K max output
GPT-5.6 Luna OpenAI	1428.5 +32.9	$0.450	Balanced −10% Summarise −34% Chat −16% Code gen −8% Agentic −39%	236K → 128K max output no video input
DeepSeek V3.2 DeepSeek	1424.6 +29	$0.290	Balanced −42% Summarise −24% Chat −59% Code gen −62% Agentic −50%	262K → 164K context 236K → 147K max output no image, video input
DeepSeek V3.2 Exp DeepSeek	1424.4 +28.8	$0.305	Balanced −39% Summarise −8% Chat −50% Code gen −58% Agentic −27%	262K → 164K context 236K → 66K max output no image, video input
Qwen3 235B A22B Instruct 2507 Qwen	1419.3 +23.7	$0.205	Balanced −59% Summarise −62% Chat −58% Code gen −57% Agentic −60%	236K → 16K max output no image, video input no extended reasoning
Qwen3 Next 80B A3B Instruct Qwen	1418.6 +23	$0.350	Balanced −30% Summarise −53% Chat −24% Code gen −18% Agentic −42%	no image, video input no extended reasoning
DeepSeek V3.1 Terminus DeepSeek	1419.6 +24	$0.453	Balanced −9% Summarise −11% Chat −20% Code gen −18% Agentic −25%	262K → 164K context 236K → 33K max output no image, video input
Step 3.5 Flash StepFun	1403.8 tie	$0.150	Balanced −70% Summarise −63% Chat −72% Code gen −74% Agentic −68%	236K → 66K max output no image, video input
Qwen3.5-Flash Qwen	1397.6 tie	$0.114	Balanced −77% Summarise −75% Chat −78% Code gen −79% Agentic −76%	236K → 66K max output
Gemini 3.7 Flash Google	1490.2 +94.6	$0.399 agentic	Agentic under 1% less	236K → 66K max output
MiMo-V2.5-Pro Xiaomi	1465 +69.4	$0.480 chat	Chat −26% Code gen −22% Agentic −39%	236K → 131K max output no image, video input
Qwen3.7 Plus Qwen	1456.2 +60.6	$0.295 summarise	Summarise −2% Chat −4% Agentic −22%	236K → 131K max output no video input
Qwen3 VL 235B A22B Instruct Qwen	1420.9 +25.3	$0.263 summarise	Summarise −12% Agentic −1%	236K → 33K max output no video input no extended reasoning
Gemini 3.1 Flash Lite Preview Google	1414.8 +19.2	$0.248 summarise	Summarise −17% Agentic −24%	236K → 66K max output
Inkling Small Thinking Machines	1411.7 +16.1	$0.645 chat	Chat −1% Agentic −11%	no video input
Qwen3.5-27B Qwen	1407.9 +12.3	$0.263 summarise	Summarise −12% Agentic under 1% less	236K → 66K max output
MiniMax M2.7 MiniMax	1405.3 tie	$0.277 summarise	Summarise −8% Chat −10% Code gen −3% Agentic −27%	262K → 205K context 236K → 131K max output no image, video input

What this compares, and what it leaves out

Quality is LMArena Elo, used under CC BY 4.0 from the official dataset. The 95% confidence interval on a difference between two scores is about ±10.68 points, so a gap smaller than that is marked tie rather than an improvement — see significance bands.
188 further models at this delivery mode carry a price but no independent score. They are absent from the comparison in both directions — unrated is not a zero, and an unmeasured model is neither an alternative nor a worse buy.
Retired models are never offered as an alternative, and a model only competes against its own delivery mode: batch trades latency for price, so it is not a like-for-like swap.
Nothing here measures latency, throughput, rate limits or how a model behaves on your prompts. Two models with the same index score are not interchangeable.
The models that nothing beats on both axes are on the value frontier, and every other model something cheaper beats is listed here.