GLM 5.3 Flash in fedi-index
The score in every slice, what it is made of, value, local run and history.
fedi-index · General
A 1,448 ±6
Percentile 89 · #33
People prefer / Solves tasks
1,459 / 1,432
Human votes vs benchmarks, Elo-eq
Value
+105◆
$0.2375 per 1M tokens, median of 58 providers · vendor $0.2375
On your hardware
- 128 GB unified memory #34 · UD-Q2_K_XL · strong compression · ≈20 tok/s
Score decomposition
A General 1,448 [1,441–1,454] Percentile 89 · #33 · components: 8
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| Epoch AI |
eci
|
152 ±1.10 | 1,453 ±9 | 36.8% | 2026-08-20 | — |
| LMArena |
text/overall
|
1,470 ±2.71 | 1,470 ±3 | 15.9% | 2026-10-02 | — |
| LMArena |
text/hard_prompts
|
1,491 ±3.17 | 1,477 ±3 | 15.8% | 2026-10-02 | — |
| LMArena |
text/instruction_following
|
1,478 ±3.81 | 1,487 ±4 | 7.7% | 2026-10-02 | — |
| LMArena |
text/expert
|
1,512 ±6.29 | 1,486 ±5 | 7.3% | 2026-10-02 | — |
| LMArena |
text/longer_query
|
1,483 ±3.61 | 1,478 ±3 | 3.9% | 2026-10-02 | — |
| LMArena |
text/multi_turn
|
1,467 ±5.30 | 1,460 ±5 | 3.7% | 2026-10-02 | — |
| Epoch AI |
gpqa_diamond
decay ×0.26 |
0.90 ±0.02 | 1,459 ±8 | 1.7% | 2026-08-26 | max |
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 1,452 [1,449–1,456] Percentile 94 · #17 · components: 2 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
webdev/overall
|
1,616 ±3.85 | 1,482 ±1 | 46.4% | 2026-10-01 | — |
| LMArena |
text/coding
|
1,506 ±4.22 | 1,476 ±4 | 43.9% | 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 Agents 1,430 [1,428–1,432] Percentile 49 · #30 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
agent/overall
|
-0.00 ±0.00 | 1,453 ±1 | 90.5% | 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
A Math 1,443 [1,434–1,452] Percentile 89 · #28 · components: 3
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/math
|
1,501 ±9.03 | 1,490 ±9 | 47.8% | 2026-10-02 | — |
| Epoch AI |
frontiermath_t1_3
decay ×0.86 |
0.56 ±0.03 | 1,442 ±6 | 31.1% | 2026-08-27 | max |
| Epoch AI |
frontiermath_t4
decay ×0.86 |
0.17 ±0.06 | 1,420 ±10 | 13.1% | 2026-08-27 | max |
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 1,460 [1,454–1,466] Percentile 88 · #14 · components: 3 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
vision/overall
|
1,301 ±4.64 | 1,476 ±4 | 64.2% | 2026-10-02 | — |
| LMArena |
vision/ocr
|
1,308 ±5.26 | 1,479 ±5 | 15.5% | 2026-10-02 | — |
| LMArena |
vision/diagram
|
1,323 ±8.61 | 1,484 ±8 | 13.7% | 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
A Writing 1,455 [1,449–1,462] Percentile 90 · #29 · components: 2 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/creative_writing
|
1,441 ±4.76 | 1,470 ±5 | 61.4% | 2026-10-02 | — |
| LMArena |
text/longer_query
|
1,483 ±3.61 | 1,478 ±3 | 32.3% | 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
A Language: 繁體中文 1,457 [1,446–1,468] Percentile 93 · #17 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/chinese
|
1,528 ±7.70 | 1,475 ±6 | 93.2% | 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
B Language: Deutsch 1,428 [1,404–1,451] Percentile 78 · #33 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/german
|
1,458 ±15 | 1,452 ±13 | 90.1% | 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
A Language: English 1,449 [1,442–1,456] Percentile 89 · #30 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/english
|
1,473 ±3.64 | 1,465 ±4 | 93.8% | 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
A Language: Español 1,441 [1,423–1,459] Percentile 84 · #24 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/spanish
|
1,469 ±11 | 1,462 ±10 | 91.7% | 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
A Language: Français 1,454 [1,435–1,472] Percentile 92 · #12 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/french
|
1,498 ±11 | 1,477 ±10 | 91.5% | 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
B Language: 日本語 1,423 [1,402–1,444] Percentile 76 · #32 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/japanese
|
1,425 ±16 | 1,445 ±12 | 90.8% | 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
A Language: 한국어 1,441 [1,420–1,462] Percentile 84 · #23 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/korean
|
1,439 ±16 | 1,465 ±12 | 90.8% | 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
B Language: Polski 1,417 [1,394–1,440] Percentile 61 · #52 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/polish
|
1,439 ±15 | 1,440 ±13 | 90.2% | 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
A Language: Русский 1,453 [1,444–1,462] Percentile 90 · #24 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/russian
|
1,471 ±5.59 | 1,470 ±5 | 93.5% | 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
History
The same in a table
| Model | 2026-10 |
|---|---|
| GLM 5.3 Flash | 1,448 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.