Cheaper alternatives to Qwen2.5 Coder 32B Instruct

Qwen LMArena 1230.2 $0.745/M on a balanced workload prices as of 2026-08-28

59 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, 75 score at least as high as Qwen2.5 Coder 32B Instruct and cost no more under at least one workload. 59 match or beat it on every capability we hold — context, maximum output, input modes, tool use and extended reasoning. 16 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 Qwen2.5 Coder 32B Instruct, it is not generated.

Where it holds

Workload	Qwen2.5 Coder 32B Instruct $/M	Better and cheaper	With a trade
Balanced	$0.745	43	14
Summarise	$0.677	57	16
Chat	$0.796	38	14
Code gen	$0.864	33	14
Agentic	$0.711	56	16

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	1230.2
Context window	33K
Max output	29K
Input modes	text
Tool use	no
Extended reasoning	no

A replacement has to clear every line above, not just the score. Full record for Qwen2.5 Coder 32B Instruct.

Better and cheaper, nothing given up 59

Each of these matches or beats Qwen2.5 Coder 32B Instruct 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	MiMo-V2.5-Pro Xiaomi	1465 +234.8	$0.544	Balanced −27% Summarise −51% Chat −40% Code gen −23% Agentic −66%
2	Qwen3.7 Plus Qwen	1456.2 +226	$0.560	Balanced −25% Summarise −56% Chat −21% Agentic −56%
3	Hy3 Tencent	1441.2 +211	$0.144	Balanced −81% Summarise −89% Chat −80% Code gen −74% Agentic −88%
4	DeepSeek V4 Flash 0423 DeepSeek	1431.6 +201.4	$0.101	Balanced −86% Summarise −90% Chat −88% Code gen −86% Agentic −92%
5	MiniMax M3 MiniMax	1434.8 +204.6	$0.525	Balanced −30% Summarise −59% Chat −26% Code gen −5% Agentic −59%
6	Qwen3.6 Plus Qwen	1436.8 +206.6	$0.731	Balanced −2% Summarise −40% Agentic −20%
7	GLM 4.7 Z.ai	1435.3 +205.1	$0.738	Balanced −1% Summarise −44% Agentic −42%
8	MiMo-V2.5 Xiaomi	1427.3 +197.1	$0.175	Balanced −77% Summarise −84% Chat −81% Code gen −75% Agentic −89%
9	GPT-5.6 Luna OpenAI	1428.5 +198.3	$0.450	Balanced −40% Summarise −71% Chat −31% Code gen −9% Agentic −66%
10	DeepSeek V3.2 DeepSeek	1424.6 +194.4	$0.290	Balanced −61% Summarise −66% Chat −66% Code gen −63% Agentic −72%
11	DeepSeek V3.2 Exp DeepSeek	1424.4 +194.2	$0.305	Balanced −59% Summarise −59% Chat −59% Code gen −59% Agentic −59%
12	Qwen3 Next 80B A3B Instruct Qwen	1418.6 +188.4	$0.350	Balanced −53% Summarise −79% Chat −38% Code gen −19% Agentic −67%
13	DeepSeek V3.1 Terminus DeepSeek	1419.6 +189.4	$0.453	Balanced −39% Summarise −60% Chat −34% Code gen −19% Agentic −58%
14	Qwen3 VL 235B A22B Instruct Qwen	1420.9 +190.7	$0.632	Balanced −15% Summarise −61% Agentic −44%
15	Qwen3.5-122B-A10B Qwen	1417.9 +187.7	$0.715	Balanced −4% Summarise −48% Agentic −25%
16	Gemini 3.1 Flash Lite Preview Google	1414.8 +184.6	$0.563	Balanced −24% Summarise −63% Chat −14% Agentic −57%
17	Inkling Small Thinking Machines	1411.7 +181.5	$0.637	Balanced −14% Summarise −43% Chat −19% Agentic −50%
18	Step 3.5 Flash StepFun	1403.8 +173.6	$0.150	Balanced −80% Summarise −84% Chat −77% Code gen −75% Agentic −82%
19	Qwen3.5-27B Qwen	1407.9 +177.7	$0.536	Balanced −28% Summarise −61% Chat −7% Agentic −44%
20	MiniMax M2.7 MiniMax	1405.3 +175.1	$0.525	Balanced −30% Summarise −59% Chat −26% Code gen −5% Agentic −59%
21	Qwen3.5-Flash Qwen	1397.6 +167.4	$0.114	Balanced −85% Summarise −89% Chat −82% Code gen −79% Agentic −87%
22	Qwen3.5-35B-A3B Qwen	1395.6 +165.4	$0.500	Balanced −33% Summarise −56% Chat −18% Code gen −2% Agentic −44%
23	Qwen3 30B A3B Instruct 2507 Qwen	1384.3 +154.1	$0.084	Balanced −89% Summarise −92% Chat −87% Code gen −84% Agentic −90%
24	GLM 4.5 Air Z.ai	1382.8 +152.6	$0.310	Balanced −58% Summarise −80% Chat −51% Code gen −36% Agentic −75%
25	Solar Pro 4 Upstage	1376.2 +146	$0.052	Balanced −93% Summarise −96% Chat −93% Code gen −90% Agentic −96%
26	DeepSeek V3 0324 DeepSeek	1375 +144.8	$0.438	Balanced −41% Summarise −58% Chat −31% Code gen −19% Agentic −49%
27	GLM 4.6V Z.ai	1374.7 +144.5	$0.450	Balanced −40% Summarise −62% Chat −41% Code gen −26% Agentic −66%
28	GPT-5.4 Nano OpenAI	1372.8 +142.6	$0.463	Balanced −38% Summarise −70% Chat −29% Code gen −6% Agentic −65%
29	gpt-oss-120b OpenAI	1365.6 +135.4	$0.070	Balanced −91% Summarise −94% Chat −89% Code gen −86% Agentic −92%
30	GPT-5 Mini OpenAI	1373.4 +143.2	$0.688	Balanced −8% Summarise −60% Agentic −47%
31	Qwen3 Next 80B A3B Thinking Qwen	1367.5 +137.3	$0.412	Balanced −45% Summarise −70% Chat −28% Code gen −10% Agentic −57%
32	Gemma 3 27B Google	1358.3 +128.1	$0.172	Balanced −77% Summarise −87% Chat −73% Code gen −65% Agentic −84%
33	Mercury 2 Inception	1357.8 +127.6	$0.375	Balanced −50% Summarise −69% Chat −52% Code gen −38% Agentic −73%
34	MiniMax M2.5 MiniMax	1359 +128.8	$0.472	Balanced −37% Summarise −64% Chat −35% Code gen −15% Agentic −65%
35	Trinity Large Thinking Arcee	1341.9 +111.7	$0.378	Balanced −49% Summarise −70% Chat −47% Code gen −32% Agentic −69%
36	MiniMax M2 MiniMax	1342.1 +111.9	$0.446	Balanced −40% Summarise −57% Chat −30% Code gen −17% Agentic −48%
37	GPT-4.1 Mini OpenAI	1340.5 +110.3	$0.700	Balanced −6% Summarise −45% Chat −1% Agentic −44%
38	GPT-5 Nano OpenAI	1320.3 +90.1	$0.138	Balanced −82% Summarise −92% Chat −78% Code gen −70% Agentic −89%
39	Olmo 3 32B Think Allen AI	1298.5 +68.3	$0.237	Balanced −68% Summarise −75% Chat −64% Code gen −58% Agentic −72%
40	Granite 4.1 8B IBM	1291.6 +61.4	$0.063	Balanced −92% Summarise −92% Chat −91% Code gen −91% Agentic −92%
41	gpt-oss-20b OpenAI	1287.8 +57.6	$0.055	Balanced −93% Summarise −95% Chat −91% Code gen −90% Agentic −94%
42	GPT-4.1 Nano OpenAI	1284.8 +54.6	$0.175	Balanced −77% Summarise −86% Chat −75% Code gen −68% Agentic −86%
43	Llama 3.3 70B Instruct Meta	1274.9 +44.7	$0.710	Balanced −5% Chat −11% Code gen −18% Agentic under 1% less
44	Gemini 3.7 Flash Google	1490.2 +260	$0.354 summarise	Summarise −48% Agentic −44%
45	Kimi K2.5 Moonshot AI	1445.2 +215	$0.432 summarise	Summarise −36% Agentic −31%
46	GLM 5 Z.ai	1445.2 +215	$0.529 summarise	Summarise −22% Agentic −28%
47	Qwen3.8 27B Qwen	1440.8 +210.6	$0.434 summarise	Summarise −36% Agentic −24%
48	GLM 4.6 Z.ai	1439.8 +209.6	$0.461 summarise	Summarise −32% Agentic −32%
49	DeepSeek V4 Pro 0423 DeepSeek	1439.2 +209	$0.526 agentic	Agentic −26%
50	Gemini 3.5 Flash Lite Google	1436.5 +206.3	$0.333 summarise	Summarise −51% Agentic −34%
51	Qwen3.5 397B A17B Qwen	1438.3 +208.1	$0.487 summarise	Summarise −28% Agentic −4%
52	GLM 4.5 Z.ai	1429.4 +199.2	$0.540 summarise	Summarise −20% Agentic −23%
53	R1 0528 DeepSeek	1427.9 +197.7	$0.540 summarise	Summarise −20% Agentic −7%
54	Gemini 2.5 Flash Google	1417.3 +187.1	$0.333 summarise	Summarise −51% Agentic −34%
55	DeepSeek V3.1 DeepSeek	1419.1 +188.9	$0.605 summarise	Summarise −11%
56	Qwen3 235B A22B Thinking 2507 Qwen	1413.8 +183.6	$0.334 summarise	Summarise −51% Agentic −24%
57	Qwen3 VL 235B A22B Thinking Qwen	1400.6 +170.4	$0.580 summarise	Summarise −14%
58	Nova 2 Lite Amazon	1362.2 +132	$0.410 summarise	Summarise −39% Agentic −11%
59	MiniMax M1 MiniMax	1341.9 +111.7	$0.632 summarise	Summarise −7%

Cheaper and higher-scoring, but you give something up 16

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 +211.5	$0.152	Balanced −80% Summarise −87% Chat −78% Code gen −73% Agentic −85%	29K → 16K max output
Gemma 4 26B A4B Google	1434.6 +204.4	$0.138	Balanced −82% Summarise −88% Chat −78% Code gen −73% Agentic −84%	29K → 16K max output
Qwen3 235B A22B Instruct 2507 Qwen	1419.3 +189.1	$0.205	Balanced −72% Summarise −83% Chat −66% Code gen −58% Agentic −78%	29K → 16K max output
GLM 4.7 Flash Z.ai	1352.9 +122.7	$0.145	Balanced −81% Summarise −91% Chat −77% Code gen −70% Agentic −89%	29K → 16K max output
Qwen3 32B Qwen	1340.1 +109.9	$0.130	Balanced −83% Summarise −87% Chat −80% Code gen −77% Agentic −85%	29K → 16K max output
Gemma 3 12B Google	1334.2 +104	$0.075	Balanced −90% Summarise −92% Chat −89% Code gen −87% Agentic −91%	29K → 16K max output
DeepSeek V3 DeepSeek	1332.6 +102.4	$0.450	Balanced −40% Summarise −56% Chat −29% Code gen −17% Agentic −48%	29K → 16K max output
Qwen3 30B A3B Qwen	1316.9 +86.7	$0.215	Balanced −71% Summarise −79% Chat −66% Code gen −60% Agentic −75%	29K → 16K max output
Gemma 3 4B Google	1290.8 +60.6	$0.063	Balanced −92% Summarise −92% Chat −91% Code gen −91% Agentic −92%	29K → 16K max output
GPT-4o-mini (2024-07-18) OpenAI	1286.6 +56.4	$0.263	Balanced −65% Summarise −78% Chat −61% Code gen −52% Agentic −76%	29K → 16K max output
Qwen2.5 72B Instruct Qwen	1269.1 +38.9	$0.370	Balanced −50% Summarise −47% Chat −53% Code gen −56% Agentic −49%	29K → 16K max output
Llama 3.1 70B Instruct Meta	1261 +30.8	$0.400	Balanced −46% Summarise −41% Chat −50% Code gen −54% Agentic −44%	29K → 16K max output
Mistral Small 3 Mistral	1233.6 tie	$0.058	Balanced −92% Summarise −92% Chat −92% Code gen −92% Agentic −92%	29K → 16K max output
Gemma 2 27B Google	1231.5 tie	$0.650	Balanced −13% Summarise −4% Chat −18% Code gen −25% Agentic −9%	33K → 8K context 29K → 2K max output
Qwen3 235B A22B Qwen	1366 +135.8	$0.523 summarise	Summarise −23% Agentic −7%	29K → 8K max output
GLM 4.5V Z.ai	1333.6 +103.4	$0.520 summarise	Summarise −23% Agentic −31%	29K → 16K max output

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.