Granite 4.2 8B in fedi-index
The score in every slice, what it is made of, value, local run and history.
fedi-index · General
C 1,310 ±6
Percentile 37 · #186 · provisional
People prefer / Solves tasks
1,318 / —
Human votes vs benchmarks, Elo-eq
Value
-14
$0.11 per 1M tokens, median of 4 providers
On your hardware
- GPU 12 GB #11 · Q6_K_L · ≈34 tok/s
- GPU 24 GB #17 · Q8_0 · ≈75 tok/s
- Mac 64 GB #15 · Q8_0 · ≈38 tok/s
- 128 GB unified memory #30 · Q8_0 · ≈18 tok/s
- CPU, 32 GB RAM #14 · Q4_K_M · ≈5 tok/s
Score decomposition
C General 1,310 [1,304–1,316] Percentile 37 · #186 · components: 6 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/overall
|
1,318 ±5.66 | 1,318 ±6 | 28.0% | 2026-10-02 | — |
| LMArena |
text/hard_prompts
|
1,327 ±7.10 | 1,329 ±6 | 27.0% | 2026-10-02 | — |
| LMArena |
text/instruction_following
|
1,299 ±9.54 | 1,317 ±9 | 11.7% | 2026-10-02 | — |
| LMArena |
text/expert
|
1,382 ±17 | 1,380 ±14 | 8.7% | 2026-10-02 | — |
| LMArena |
text/longer_query
|
1,322 ±8.76 | 1,327 ±8 | 6.2% | 2026-10-02 | — |
| LMArena |
text/multi_turn
|
1,299 ±14 | 1,307 ±13 | 4.5% | 2026-10-02 | — |
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 1,342 [1,325–1,358] Percentile 40 · #166 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/coding
|
1,378 ±11 | 1,362 ±10 | 87.2% | 2026-10-02 | — |
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 1,268 [1,250–1,286] Percentile 31 · #185 · components: 2 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/creative_writing
|
1,233 ±14 | 1,258 ±14 | 64.9% | 2026-10-02 | — |
| LMArena |
text/longer_query
outlier |
1,322 ±8.76 | 1,327 ±8 | 23.6% | 2026-10-02 | — |
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 1,311 [1,294–1,327] Percentile 37 · #172 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/english
|
1,336 ±8.51 | 1,320 ±9 | 92.2% | 2026-10-02 | — |
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 1,289 [1,262–1,316] Percentile 27 · #172 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/russian
|
1,281 ±17 | 1,300 ±16 | 88.6% | 2026-10-02 | — |
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 |
|---|---|
| Granite 4.2 8B | 1,310 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.