Undominated.ai

Cheaper alternatives to Qwen3.5 397B A17B

Qwen LMArena 1438.3 $0.877/M on a balanced workload prices as of 2026-08-28

2 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, 10 score at least as high as Qwen3.5 397B A17B and cost no more under at least one workload. 2 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 Qwen3.5 397B A17B, it is not generated.

Where it holds

WorkloadQwen3.5 397B A17B $/MBetter and cheaperWith a trade
Balanced$0.87715
Summarise$0.48726
Chat$1.1718
Code gen$1.5618
Agentic$0.68228

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

Intelligence1438.3
Context window262K
Max output66K
Input modestext, image, video
Tool useyes
Extended reasoningyes

A replacement has to clear every line above, not just the score. Full record for Qwen3.5 397B A17B.

Better and cheaper, nothing given up 2

Each of these matches or beats Qwen3.5 397B A17B on context, maximum output, input modes, tool use and reasoning, scores at least as high, and costs no more.

#ModelIntelligence$/MHolds under
1Gemini 3.7 Flash Google1490.2 +51.9$0.750
  • Balanced −15%
  • Summarise −27%
  • Chat −25%
  • Code gen −20%
  • Agentic −42%
2Qwen3.8 27B Qwen1440.8 tie$0.434 summarise
  • Summarise −11%
  • Agentic −21%

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 +26.7$0.544
  • Balanced −38%
  • Summarise −32%
  • Chat −59%
  • Code gen −58%
  • Agentic −64%
  • no image, video input
Qwen3.7 Plus Qwen1456.2 +17.9$0.560
  • Balanced −36%
  • Summarise −39%
  • Chat −46%
  • Code gen −44%
  • Agentic −54%
  • no video input
Gemma 4 31B Google1441.7 tie$0.152
  • Balanced −83%
  • Summarise −81%
  • Chat −85%
  • Code gen −85%
  • Agentic −85%
  • 66K → 16K max output
Hy3 Tencent1441.2 tie$0.144
  • Balanced −84%
  • Summarise −84%
  • Chat −86%
  • Code gen −86%
  • Agentic −88%
  • no image, video input
GLM 4.6 Z.ai1439.8 tie$0.875
  • Balanced under 1% less
  • Summarise −5%
  • Chat −16%
  • Code gen −12%
  • Agentic −29%
  • 262K → 205K context
  • no image, video input
Kimi K2.5 Moonshot AI1445.2 tie$0.432 summarise
  • Summarise −11%
  • Chat −10%
  • Code gen −4%
  • Agentic −28%
  • no video input
GLM 5 Z.ai1445.2 tie$0.984 chat
  • Chat −16%
  • Code gen −13%
  • Agentic −25%
  • 262K → 205K context
  • no image, video input
DeepSeek V4 Pro 0423 DeepSeek1439.2 tie$0.979 chat
  • Chat −16%
  • Code gen −15%
  • Agentic −23%
  • 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.