Undominated.ai

Cheaper alternatives to Qwen3 VL 235B A22B Instruct

Qwen LMArena 1420.9 $0.632/M on a balanced workload prices as of 2026-08-28

4 models are 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, 12 score at least as high as Qwen3 VL 235B A22B Instruct and cost no more under at least one workload. 4 match or beat it on every capability we hold — context, maximum output, input modes, tool use and extended reasoning. 8 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 VL 235B A22B Instruct, it is not generated.

Where it holds

WorkloadQwen3 VL 235B A22B Instruct $/MBetter and cheaperWith a trade
Balanced$0.63247
Summarise$0.26325
Chat$0.85348
Code gen$1.2248
Agentic$0.39847

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

Intelligence1420.9
Context window262K
Max output33K
Input modestext, image
Tool useyes
Extended reasoningno

A replacement has to clear every line above, not just the score. Full record for Qwen3 VL 235B A22B Instruct.

Better and cheaper, nothing given up 4

Each of these matches or beats Qwen3 VL 235B A22B Instruct on context, maximum output, input modes, tool use and reasoning, scores at least as high, and costs no more.

#ModelIntelligence$/MHolds under
1Qwen3.7 Plus Qwen1456.2 +35.3$0.560
  • Balanced −11%
  • Chat −26%
  • Code gen −28%
  • Agentic −22%
2MiniMax M3 MiniMax1434.8 +13.9$0.525
  • Balanced −17%
  • Chat −31%
  • Code gen −32%
  • Agentic −27%
3MiMo-V2.5 Xiaomi1427.3 tie$0.175
  • Balanced −72%
  • Summarise −59%
  • Chat −82%
  • Code gen −82%
  • Agentic −80%
4GPT-5.6 Luna OpenAI1428.5 tie$0.450
  • Balanced −29%
  • Summarise −24%
  • Chat −36%
  • Code gen −35%
  • Agentic −39%

Cheaper and higher-scoring, but you give something up 8

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.

ModelIntelligence$/MHolds underWhat you give up
MiMo-V2.5-Pro Xiaomi1465 +44.1$0.544
  • Balanced −14%
  • Chat −44%
  • Code gen −46%
  • Agentic −39%
  • no image input
Gemma 4 31B Google1441.7 +20.8$0.152
  • Balanced −76%
  • Summarise −65%
  • Chat −79%
  • Code gen −81%
  • Agentic −74%
  • 33K → 16K max output
Hy3 Tencent1441.2 +20.3$0.144
  • Balanced −77%
  • Summarise −71%
  • Chat −81%
  • Code gen −81%
  • Agentic −79%
  • no image input
Gemma 4 26B A4B Google1434.6 +13.7$0.138
  • Balanced −78%
  • Summarise −68%
  • Chat −79%
  • Code gen −81%
  • Agentic −72%
  • 33K → 16K max output
DeepSeek V4 Flash 0423 DeepSeek1431.6 +10.7$0.101
  • Balanced −84%
  • Summarise −75%
  • Chat −89%
  • Code gen −90%
  • Agentic −86%
  • no image input
DeepSeek V3.2 DeepSeek1424.6 tie$0.290
  • Balanced −54%
  • Summarise −13%
  • Chat −68%
  • Code gen −74%
  • Agentic −50%
  • 262K → 164K context
  • no image input
DeepSeek V3.2 Exp DeepSeek1424.4 tie$0.305
  • Balanced −52%
  • Chat −62%
  • Code gen −71%
  • Agentic −27%
  • 262K → 164K context
  • no image input
GLM 4.7 Z.ai1435.3 +14.4$0.844 chat
  • Chat −1%
  • Code gen −3%
  • 262K → 205K context
  • no image 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.