Gemini 3.5 Flash in fedi-index
The score in every slice, what it is made of, value, local run and history.
fedi-index · General
A 1,457 ±5
Percentile 93 · #22
People prefer / Solves tasks
1,460 / 1,453
Human votes vs benchmarks, Elo-eq
Value
+49
$3.38 per 1M tokens, median of 29 providers · vendor $3.38
On your hardware
Closed weights: cloud only.
Score decomposition
A General 1,457 [1,452–1,462] Percentile 93 · #22 · components: 9
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| Epoch AI |
eci
|
154 ±1.02 | 1,473 ±8 | 32.8% | 2026-05-19 | — |
| LMArena |
text/overall
|
1,482 ±1.98 | 1,482 ±2 | 15.8% | 2026-10-02 | high |
| LMArena |
text/hard_prompts
|
1,488 ±2.33 | 1,474 ±2 | 15.8% | 2026-10-02 | high |
| LMArena |
text/instruction_following
|
1,467 ±2.85 | 1,476 ±3 | 7.8% | 2026-10-02 | high |
| LMArena |
text/expert
|
1,495 ±4.36 | 1,472 ±4 | 7.6% | 2026-10-02 | high |
| LMArena |
text/longer_query
|
1,482 ±2.71 | 1,477 ±2 | 3.9% | 2026-10-02 | high |
| Epoch AI |
simpleqa_verified
decay ×0.61 |
0.66 ±0.01 | 1,505 ±5 | 3.9% | 2026-08-27 | high |
| LMArena |
text/multi_turn
|
1,481 ±3.77 | 1,472 ±3 | 3.8% | 2026-10-02 | high |
| Epoch AI |
gpqa_diamond
decay ×0.26 |
0.93 ±0.02 | 1,471 ±8 | 1.5% | 2026-05-22 | high |
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
A Coding 1,439 [1,435–1,442] Percentile 89 · #31 · components: 3
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
webdev/overall
|
1,499 ±3.69 | 1,449 ±1 | 39.4% | 2026-10-01 | high |
| LMArena |
text/coding
|
1,492 ±3.09 | 1,464 ±3 | 38.3% | 2026-10-02 | high |
| Epoch AI |
swe_bench_verified
decay ×0.37 |
0.79 ±0.02 | 1,480 ±10 | 14.0% | 2026-06-01 | high |
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
A Math 1,454 [1,446–1,461] Percentile 92 · #20 · components: 3
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/math
|
1,504 ±6.37 | 1,493 ±6 | 51.7% | 2026-10-02 | high |
| Epoch AI |
frontiermath_t1_3
decay ×0.86 |
0.63 ±0.03 | 1,457 ±6 | 29.6% | 2026-06-10 | high |
| Epoch AI |
frontiermath_t4
decay ×0.86 |
0.27 ±0.07 | 1,435 ±11 | 11.2% | 2026-06-10 | high |
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
A Vision 1,467 [1,463–1,472] Percentile 93 · #9 · components: 3 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
vision/overall
|
1,310 ±3.61 | 1,484 ±3 | 63.3% | 2026-10-02 | high |
| LMArena |
vision/ocr
|
1,316 ±3.79 | 1,487 ±4 | 15.6% | 2026-10-02 | medium |
| LMArena |
vision/diagram
|
1,326 ±5.69 | 1,486 ±5 | 14.9% | 2026-10-02 | medium |
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
A Writing 1,474 [1,470–1,479] Percentile 96 · #13 · components: 2 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/creative_writing
|
1,470 ±3.76 | 1,500 ±4 | 61.8% | 2026-10-02 | medium |
| LMArena |
text/longer_query
|
1,482 ±2.71 | 1,477 ±2 | 32.0% | 2026-10-02 | high |
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
A Language: 繁體中文 1,457 [1,449–1,465] Percentile 93 · #16 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/chinese
|
1,526 ±5.35 | 1,474 ±4 | 93.7% | 2026-10-02 | high |
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
A Language: Deutsch 1,460 [1,442–1,478] Percentile 96 · #7 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/german
|
1,494 ±12 | 1,483 ±10 | 91.7% | 2026-10-02 | medium |
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
A Language: English 1,456 [1,451–1,462] Percentile 94 · #18 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/english
|
1,480 ±2.64 | 1,473 ±3 | 93.9% | 2026-10-02 | high |
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
A Language: Español 1,452 [1,438–1,466] Percentile 94 · #10 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/spanish
|
1,479 ±8.55 | 1,472 ±8 | 92.6% | 2026-10-02 | high |
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
A Language: Français 1,448 [1,434–1,463] Percentile 87 · #18 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/french
|
1,488 ±8.56 | 1,468 ±8 | 92.6% | 2026-10-02 | high |
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
A Language: 日本語 1,468 [1,451–1,485] Percentile 94 · #8 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/japanese
|
1,486 ±13 | 1,491 ±9 | 92.0% | 2026-10-02 | high |
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
A Language: 한국어 1,454 [1,439–1,469] Percentile 92 · #12 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/korean
|
1,452 ±11 | 1,475 ±8 | 92.5% | 2026-10-02 | medium |
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
A Language: Polski 1,485 [1,468–1,502] Percentile 100 · #1 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/polish
|
1,521 ±11 | 1,510 ±9 | 92.0% | 2026-10-02 | high |
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
A Language: Русский 1,472 [1,465–1,479] Percentile 95 · #12 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/russian
|
1,492 ±4.40 | 1,490 ±4 | 93.7% | 2026-10-02 | high |
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 |
|---|---|
| Gemini 3.5 Flash | 1,457 A |
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.