Trinity Large Thinking in fedi-index
The score in every slice, what it is made of, value, local run and history.
fedi-index · General
C 1,326 ±2
Percentile 40 · #180 · provisional
People prefer / Solves tasks
1,333 / —
Human votes vs benchmarks, Elo-eq
Value
-20
$0.40 per 1M tokens, median of 2 providers · vendor $0.3875
On your hardware
- 128 GB unified memory #44 · IQ2_XXS · strong compression · ≈31 tok/s
Score decomposition
C General 1,326 [1,324–1,328] Percentile 40 · #180 · components: 6 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/overall
|
1,340 ±2.37 | 1,340 ±2 | 25.8% | 2026-11-19 | — |
| LMArena |
text/hard_prompts
|
1,342 ±2.85 | 1,342 ±3 | 25.7% | 2026-11-19 | — |
| LMArena |
text/instruction_following
|
1,326 ±3.49 | 1,343 ±3 | 12.6% | 2026-11-19 | — |
| LMArena |
text/expert
|
1,359 ±5.98 | 1,361 ±5 | 11.9% | 2026-11-19 | — |
| LMArena |
text/longer_query
|
1,318 ±3.42 | 1,324 ±3 | 6.3% | 2026-11-19 | — |
| LMArena |
text/multi_turn
|
1,320 ±4.55 | 1,327 ±4 | 6.1% | 2026-11-19 | — |
Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology
C Coding 1,341 [1,337–1,345] Percentile 37 · #183 · components: 2 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
webdev/overall
|
1,237 ±10 | 1,375 ±3 | 44.5% | 2026-11-20 | — |
| LMArena |
text/coding
|
1,357 ±3.81 | 1,343 ±3 | 43.9% | 2026-11-19 | — |
Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology
C Math 1,339 [1,326–1,351] Percentile 40 · #157 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/math
|
1,365 ±7.63 | 1,360 ±7 | 86.6% | 2026-11-19 | — |
Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology
C Writing 1,323 [1,316–1,329] Percentile 43 · #154 · components: 2 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/creative_writing
|
1,308 ±4.79 | 1,334 ±5 | 61.2% | 2026-11-19 | — |
| LMArena |
text/longer_query
|
1,318 ±3.42 | 1,324 ±3 | 32.4% | 2026-11-19 | — |
Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology
C Language: 繁體中文 1,342 [1,332–1,353] Percentile 35 · #144 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/chinese
|
1,373 ±7.23 | 1,352 ±6 | 93.3% | 2026-11-19 | — |
Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology
C Language: Deutsch 1,347 [1,325–1,370] Percentile 32 · #101 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/german
|
1,354 ±15 | 1,363 ±13 | 90.3% | 2026-11-19 | — |
Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology
C Language: English 1,332 [1,326–1,338] Percentile 40 · #164 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/english
|
1,355 ±3.09 | 1,341 ±3 | 93.9% | 2026-11-19 | — |
Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology
D Language: Español 1,321 [1,303–1,339] Percentile 12 · #134 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/spanish
|
1,328 ±11 | 1,332 ±10 | 91.8% | 2026-11-19 | — |
Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology
D Language: Français 1,346 [1,328–1,363] Percentile 18 · #117 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/french
|
1,369 ±10 | 1,358 ±10 | 91.9% | 2026-11-19 | — |
Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology
C Language: 한국어 1,350 [1,332–1,368] Percentile 26 · #103 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/korean
|
1,306 ±13 | 1,364 ±10 | 91.6% | 2026-11-19 | — |
Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology
D Language: Polski 1,318 [1,298–1,339] Percentile 7 · #124 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/polish
|
1,309 ±13 | 1,330 ±11 | 91.0% | 2026-11-19 | — |
Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology
C Language: Русский 1,342 [1,332–1,351] Percentile 37 · #148 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/russian
|
1,338 ±5.62 | 1,351 ±5 | 93.5% | 2026-11-19 | — |
Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology
History
The same in a table
| Model | 2026-10 | 2026-11 |
|---|---|---|
| Trinity Large Thinking | 1,326 C | 1,326 C |
Badge
Show the model's fedi-index on your site or in a README. The badge updates itself; it links to this page.
Data CC BY 4.0: keep the link to fedi.software.