Arcee AITrinity Large
Trinity Large is an Arcee AI open-weights reasoning model, released 1 Apr 2026. Its best configuration (thinking reasoning effort) ranks #467 of 785 on the BenchLeader Index at 46.9 ±4.5, in the lower half. The ± is the point: 312 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in instruction following (57) and lowest in coding (37). At $0.388 per million tokens blended it is cheaper than most ranked models. Output speed of 357 tokens per second puts it in the fastest quarter, with a first answer in 6.9 s. It has been measured at 2 reasoning-effort settings; this summary describes the best-scoring one, and the tabs above switch between them. Last measured 22 Oct 2026.
- Blended price
- $0.388/M
- $0.250 in · $0.800 out
- Output speed
- 357 tok/s
- measured by Artificial Analysis
- First answer
- 6.89 s
- first token 0.60 s
- Context
- 512k
- Full answer
- 8.29 s
- median, reasoning included
- Cost per run
- $0.119
- one full Intelligence Index run
- Released
- 1 Apr 2026
- Overall index47
- Reasoning46
- Coding37
- Agents & tools40
- Maths56
- Knowledge42
- Instruction following57
- Human preference53
- Long context43
- Composite42
Reasoning-effort configurations
The same model behaves differently depending on how much it is allowed to think. Each row is one setting, scored only on the benchmarks that were run at that setting. “Not stated” collects results from publishers that did not say which setting they used; for a reasoning model that is usually its thinking mode, but we do not assume it. Pick a setting here or at the top of the page to see its price, speed and category scores.
| Effort | Index | Rank | Speed | First answer | Chat reply cost | Categories |
|---|---|---|---|---|---|---|
| thinkingbest | 46.9 | #467 | 357 tok/s | 6.89 s | $0.0003 | Agents & tools 40 · Coding 37 · Composite 42 · Human preference 53 · Instruction following 57 · Knowledge 42 · Long context 43 · Maths 56 · Reasoning 46 |
| not stated | – | – | 180 tok/s | 0.60 s | – | Coding 57 · Human preference 55 · Maths 53 · Reasoning 56 |
How its index has moved
8 Sept 2026 to 23 Oct 2026The index is recomputed from scratch every day, so a line moves when a new benchmark result lands, when a publisher revises a score, or when the models it is normalised against change. Early movement usually means the score is still settling.
Benchmark results
See all LLM benchmarks. One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.
Reasoning
| Benchmark | thinking | not stated | Source | Trend |
|---|---|---|---|---|
| Arena Hard Prompts | 1385#182 | 1403#170 | Arena (LMArena) | |
| GPQA Diamond (AA)not in index | 75.2%#251 | – | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | 15.8%#249 | – | Artificial Analysis | |
| CritPt | 0.9%#237 | – | Artificial Analysis |
Coding
| Benchmark | thinking | not stated | Source | Trend |
|---|---|---|---|---|
| SciCode | 40.6%#157 | – | SciCode | |
| Arena Coding | 1413#185 | 1443#153 | Arena (LMArena) | |
| Arena WebDev | 1238#129 | – | Arena (LMArena) | |
| SciCode (AA)not in index | 40.6%#169 | – | Artificial Analysis | |
| Surface Evolver Benchnot in index | 15.6%#28 | – | Epoch AI Benchmarking Hub | |
| Terminal-Bench 4.0 (AA)not in index | 0.5%#139 | – | Artificial Analysis | |
| Terminal-Bench 2.1 (AA)not in index | 20.6%#174 | – | Artificial Analysis |
Agents & tools
| Benchmark | thinking | Source | Trend |
|---|---|---|---|
| Terminal-Bench Hard | 22.7%#153 | Artificial Analysis | |
| τ²-Bench Telecom (AA)not in index | 90.1%#65 | Artificial Analysis | |
| GDPval-AA v2.1not in index | 0.0%#124 | Artificial Analysis | |
| τ³-Banking (AA) | 5.8%#173 | Artificial Analysis | |
| Terminal-Bench 2.0 (Vals)not in index | 18.0%#57 | Vals AI | |
| AutomationBenchnot in index | 1.3%#191 | Artificial Analysis | |
| GDP.pdf (AA)not in index | 1.2%#198 | Artificial Analysis |
Maths
| Benchmark | thinking | not stated | Source | Trend |
|---|---|---|---|---|
| Arena Maths | 1383#169 | 1362#192 | Arena (LMArena) |
Knowledge
| Benchmark | thinking | Source | Trend |
|---|---|---|---|
| AA-Omniscience | -44.1#346 | Artificial Analysis | |
| CorpFin | 54.7%#90 | Vals AI | |
| CaseLaw v2not in index | 57.9%#29 | Vals AI | |
| AA-Omniscience: accuracynot in index | 22.5%#274 | Artificial Analysis | |
| AA-Omniscience: non-hallucinationnot in index | 14.1%#354 | Artificial Analysis |
Instruction following
| Benchmark | thinking | not stated | Source | Trend |
|---|---|---|---|---|
| IFBench | 56.3%#144 | – | Artificial Analysis | |
| Arena Instruction Followingnot in index | 1359#182 | 1370#172 | Arena (LMArena) |
Human preference
| Benchmark | thinking | not stated | Source | Trend |
|---|---|---|---|---|
| Arena Text | 1367#184 | 1379#175 | Arena (LMArena) | |
| Arena Creative Writingnot in index | 1333#185 | 1352#166 | Arena (LMArena) | |
| Arena Multi-turnnot in index | 1346#198 | 1377#174 | Arena (LMArena) | |
| Arena Longer Queriesnot in index | 1365#192 | 1391#169 | Arena (LMArena) |
Long context
| Benchmark | thinking | Source | Trend |
|---|---|---|---|
| AA-LCR | 38.0%#365 | Artificial Analysis |
Composite
| Benchmark | thinking | Source | Trend |
|---|---|---|---|
| AA Intelligence Index v4.3.2 | 10.8#371 | Artificial Analysis |
Where to run it
Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.
| Provider | Speed | First token | Input $/M | Output $/M | Context | Quantisation |
|---|---|---|---|---|---|---|
| Arcee AI | 200 tok/s | 0.60 s | $0.250 | $0.800 | 262k | – |
Price history
The price above, per 1M tokens, every day it changed.
What a task costs
Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0003 | – | 7.7 s |
| Summarize a 30-page report | 12,000 / 600 | $0.0035 | – | 8.6 s |
| Code edit | 6,000 / 1,500 | $0.0027 | – | 11.1 s |
| Agentic coding session | 60,000 / 4,000 | $0.018 | – | 18.1 s |
| Structured extraction | 2,000 / 200 | $0.0007 | – | 7.5 s |
See also
Questions people ask
- What are Trinity Large's benchmark scores?
- As of 24 Oct 2026, Trinity Large (thinking effort) scores 46.9 ±4.5 on the BenchLeader Index, strongest in instruction following (57) and weakest in coding (37). That rests on 14 independent benchmark results, and every score on this page links to its publisher.
- How does Trinity Large rank?
- Trinity Large (thinking effort) is #467 of 785 ranked configurations on 24 Oct 2026. 312 other configurations score within its margin, so the exact position matters less than the band. The leader is Claude Opus 5.5 (max effort) at 72.6.
- How much does Trinity Large cost per 1M tokens?
- As of 24 Oct 2026, Trinity Large costs $0.250 per 1M input tokens and $0.800 per 1M output tokens, $0.388 blended (three input tokens per output token). The list price last changed on 24 Sept 2026.
- When was Trinity Large released?
- Arcee AI released Trinity Large on 1 Apr 2026. BenchLeader has results for it at 2 reasoning-effort settings.
Data as of 24 Oct 2026. Compare these configurations.
Cite as: BenchLeader, “Trinity Large: benchmarks, pricing, speed and rank”, data as of 24 Oct 2026, https://www.benchleader.com/models/trinity-large.