Alternatives to Trinity Large Thinking

Arcee LMArena 1340 $0.388/M on a balanced workload prices as of 2026-10-06

11 models are higher-scoring or cheaper, with neither dimension worse.

Of the 144 models carrying both an independent score and a published price at the same delivery mode, 28 score at least as high as Trinity Large Thinking and cost no more under at least one listed workload, with at least one of those dimensions improved. 11 match or beat its recorded context, maximum output, input modes, tool use and extended reasoning. 17 have a recorded capability loss, named on every row.

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 only while a candidate improves score or price without worsening the other under a listed workload. When that comparison no longer holds, it is not generated.

Where it holds

Workload	Trinity Large Thinking $/M	Recorded capabilities preserved	Recorded capability losses
Balanced	$0.388	8	16
Summarise	$0.223	11	15
Chat	$0.413	8	14
Code gen	$0.565	8	14
Agentic	$0.219	8	14

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	1340
Context window	262K
Max output	80K
Input modes	text
Tool use	yes
Extended reasoning	yes

Preserving recorded capabilities requires matching or beating these fields. Full record for Trinity Large Thinking, or check Trinity Large Thinking against the whole catalogue.

Recorded capabilities preserved 11

Each option has a higher score or a lower price, with neither dimension worse under the listed workloads. It matches or beats Trinity Large Thinking on recorded context, maximum output, input modes, tool use and extended reasoning.

#	Model	Intelligence	$/M	Holds under
1	GLM 5.3 Flash Z.ai	1469.6 +129.6	$0.238	Balanced −39% Summarise −40% Chat −38% Code gen −38% Agentic −40%
2	DeepSeek V4.1 Flash DeepSeek	1462.5 +122.5	$0.182	Balanced −53% Summarise −70% Chat −40% Code gen −36% Agentic −52%
3	MiMo-V2.6-Flash Xiaomi	1456.4 +116.4	$0.175	Balanced −55% Summarise −52% Chat −63% Code gen −62% Agentic −64%
4	Hy3 Tencent	1440.7 +100.7	$0.231	Balanced −40% Summarise −45% Chat −37% Code gen −36% Agentic −40%
5	Gemma 4 26B A4B Google	1433.9 +93.9	$0.127	Balanced −67% Summarise −69% Chat −64% Code gen −64% Agentic −64%
6	DeepSeek V4 Flash 0423 DeepSeek	1432.1 +92.1	$0.113	Balanced −71% Summarise −67% Chat −75% Code gen −76% Agentic −72%
7	MiMo-V2.5 Xiaomi	1427.5 +87.5	$0.175	Balanced −55% Summarise −52% Chat −63% Code gen −62% Agentic −64%
8	GPT-6 Luna OpenAI	1391.5 +51.5	$0.200	Balanced −48% Summarise −58% Chat −44% Code gen −41% Agentic −51%
9	Qwen3.8 27B Qwen	1440.7 +100.7	$0.183 summarise	Summarise −18%
10	GPT-5.6 Luna OpenAI	1431 +91	$0.199 summarise	Summarise −11%
11	GPT-5.4 Nano OpenAI	1372.1 +32.1	$0.201 summarise	Summarise −10%

Alternatives with recorded capability losses 17

These improve score or price without worsening the other under the listed workloads, but reduce at least one recorded capability. Read the last column before switching.

Model	Intelligence	$/M	Holds under	Recorded capability loss
Gemma 4 31B Google	1443.4 +103.4	$0.205	Balanced −47% Summarise −31% Chat −41% Code gen −48% Agentic −18%	80K → 16K max output
DeepSeek V3.2 DeepSeek	1424.5 +84.5	$0.290	Balanced −25% Chat −35% Code gen −43% Agentic −9%	262K → 164K context 80K → 66K max output
DeepSeek V3.2 Exp DeepSeek	1422.4 +82.4	$0.305	Balanced −21% Chat −21% Code gen −37%	262K → 164K context
Qwen3 235B A22B Instruct 2507 Qwen	1419.3 +79.3	$0.205	Balanced −47% Summarise −49% Chat −34% Code gen −35% Agentic −28%	80K → 16K max output no extended reasoning
Qwen3 VL 235B A22B Instruct Qwen	1420.1 +80.1	$0.370	Balanced −5% Summarise −7%	80K → 33K max output no extended reasoning
Qwen3 Next 80B A3B Instruct Qwen	1417.3 +77.3	$0.268	Balanced −31% Summarise −41% Chat −10% Code gen −10% Agentic −9%	no extended reasoning
Step 3.5 Flash StepFun	1403 +63	$0.150	Balanced −61% Summarise −51% Chat −56% Code gen −61% Agentic −41%	80K → 66K max output
Qwen3.5-Flash Qwen	1397 +57	$0.114	Balanced −71% Summarise −67% Chat −65% Code gen −68% Agentic −57%	80K → 66K max output
Qwen3.5-35B-A3B Qwen	1394.4 +54.4	$0.355	Balanced −8% Summarise −30% Agentic −2%	80K → 33K max output
Qwen3 30B A3B Instruct 2507 Qwen	1384.1 +44.1	$0.143	Balanced −63% Summarise −55% Chat −58% Code gen −62% Agentic −45%	no extended reasoning
GLM 4.5 Air Z.ai	1383.9 +43.9	$0.310	Balanced −20% Summarise −39% Chat −6% Code gen −2% Agentic −20%	262K → 131K context
gpt-oss-120b OpenAI	1365.4 +25.4	$0.065	Balanced −83% Summarise −83% Chat −79% Code gen −80% Agentic −77%	262K → 131K context
Gemma 3 27B Google	1357.8 +17.8	$0.100	Balanced −74% Summarise −62% Chat −73% Code gen −77% Agentic −58%	262K → 131K context no extended reasoning
GLM 4.7 Flash Z.ai	1350.9 +10.9	$0.145	Balanced −63% Summarise −72% Chat −56% Code gen −54% Agentic −63%	262K → 200K context
Mercury 2 Inception	1355.2 +15.2	$0.375	Balanced −3% Summarise −6% Chat −7% Code gen −6% Agentic −13%	262K → 128K context 80K → 50K max output
Qwen3 32B Qwen	1340.1 tie	$0.130	Balanced −66% Summarise −60% Chat −61% Code gen −65% Agentic −50%	262K → 131K context 80K → 16K max output
Qwen3 Next 80B A3B Thinking Qwen	1368.3 +28.3	$0.203 summarise	Summarise −9%	80K → 33K 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 ±10.47 points, so a gap smaller than that is marked tie rather than an improvement — see significance bands.
204 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.
Models with no higher-scoring or cheaper alternative that is no worse on the other axis are on the value frontier. Models with such an alternative are listed here.

Choose models

Compare costs

Build with AI

Understand the numbers

Check our work

Reuse the data