Reasoning benchmark / ARC Prize
ARC-AGI-3
Lumina VerifiedLumina checked the published sources, versions, model labels, score extraction and conversion. External evaluations were run by their publishers; Lumina has not independently rerun or reproduced them.Verified interactive reasoning results on ARC-AGI-3.
ARC-AGI-3 · Lumina Score
Semi-Private · Standard harness
20 of 69 configurations
Lumina Score vs Evaluation Cost
Publisher-reported evaluation cost; not a per-task price. · Standard harness
15 of 24 families
Missing resource measurements remain unavailable. Missing metrics are omitted from their graph; quality scores remain available.
Results table
| Rank | |||||
|---|---|---|---|---|---|
| 1 | GPT-6 AstraMax | 1,239 | 62.71% | $26,097.5 | ARC-AGI-3 Standard |
| 2 | GPT-6 AstraxHigh | 1,224 | 59.34% | $37,317.39 | ARC-AGI-3 Standard |
| 3 | GPT-6 AstraHigh | 1,204 | 54.82% | $40,704.72 | ARC-AGI-3 Standard |
| 4 | GPT-6.1 SolMax | 1,195 | 52.73% | $7,632.48 | ARC-AGI-3 Standard |
| 5 | Gemini 4 ArgonLow | 1,182 | 49.79% | $5,820.53 | ARC-AGI-3 Standard |
| 6 | Gemini 4 ArgonMedium | 1,166 | 46.27% | $6,179.25 | ARC-AGI-3 Standard |
| 7 | GPT-6.1 SolxHigh | 1,138 | 39.93% | $9,104.43 | ARC-AGI-3 Standard |
| 8 | GPT-6 AstraMedium | 1,132 | 38.59% | $48,090.35 | ARC-AGI-3 Standard |
| 9 | GPT-6 AstraNone | 1,117 | 35.18% | $49,791.11 | ARC-AGI-3 Standard |
| 10 | Claude Opus 5High | 1,094 | 30.16% | $20,657.37 | ARC-AGI-3 Standard |
| 11 | GPT-6.1 SolHigh | 1,079 | 26.72% | $9,757.62 | ARC-AGI-3 Standard |
| 12 | Claude Sonnet 5.5xHigh | 1,068 | 24.31% | $7,394.47 | ARC-AGI-3 Standard |
| 13 | GPT-6 AstraLow | 1,038 | 17.45% | $38,166.47 | ARC-AGI-3 Standard |
| 14 | GPT-6.1 SolMedium | 1,007 | 10.57% | $7,369.78 | ARC-AGI-3 Standard |
| 15 | Gemini 3.8 FlashHigh | 1,006 | 10.37% | $4,398.28 | ARC-AGI-3 Standard |
| 16 | GPT-5.6 SolMax | 995 | 7.78% | $25,064.11 | ARC-AGI-3 Standard |
| 17 | GPT-5.6 SolxHigh | 991 | 6.99% | $19,216.38 | ARC-AGI-3 Standard |
| 18 | Gemini 3.8 FlashLow | 987 | 5.99% | $2,669.06 | ARC-AGI-3 Standard |
| 19 | GPT-6 SolMax | 981 | 4.62% | $5,554.49 | ARC-AGI-3 Standard |
| 20 | Gemini 3.8 FlashMedium | 978 | 3.99% | $2,727.66 | ARC-AGI-3 Standard |
| 20 | GPT-6.1 SolLow | 978 | 3.92% | $7,110.56 | ARC-AGI-3 Standard |
| 22 | GPT-5.6 SolHigh | 970 | 2.15% | $15,176.07 | ARC-AGI-3 Standard |
| 23 | Grok 4.6xHigh | 969 | 2.11% | $5,612 | ARC-AGI-3 Standard |
| 24 | GPT-6 SolxHigh | 968 | 1.8% | $5,105.62 | ARC-AGI-3 Standard |
| 24 | Grok 4.7High | 968 | 1.73% | $3,741.65 | ARC-AGI-3 Standard |
| 24 | Grok 4.7Medium | 968 | 1.71% | $3,795.49 | ARC-AGI-3 Standard |
| 24 | Grok 4.7xHigh | 968 | 1.82% | $2,731.58 | ARC-AGI-3 Standard |
| 28 | Claude Opus 4.8High | 967 | 1.52% | $10,000 | ARC-AGI-3 Standard |
| 29 | Claude Sonnet 5.5High | 965 | 1.2% | $5,487.23 | ARC-AGI-3 Standard |
| 29 | GPT-5.6 SolMedium | 965 | 1.07% | $12,971.17 | ARC-AGI-3 Standard |
| 31 | DeepSeek V4.1 FlashMax | 964 | 0.85% | $1,164.91 | ARC-AGI-3 Standard |
| 31 | GPT-5.6 TerraMax | 964 | 0.8% | $7,918.31 | ARC-AGI-3 Standard |
| 31 | GPT-6 SolHigh | 964 | 0.9% | $5,072.45 | ARC-AGI-3 Standard |
| 34 | GPT-5.6 TerraxHigh | 963 | 0.65% | $6,804.04 | ARC-AGI-3 Standard |
| 35 | Claude Haiku 5.5Max | 962 | 0.45% | $388.11 | ARC-AGI-3 Standard |
| 35 | Claude Sonnet 5.5Low | 962 | 0.4% | $4,756.68 | ARC-AGI-3 Standard |
| 35 | Claude Sonnet 5.5Medium | 962 | 0.4% | $4,839.68 | ARC-AGI-3 Standard |
| 35 | Claude Opus 4.6Max | 962 | 0.51% | $8,866.2 | ARC-AGI-3 Standard |
| 35 | DeepSeek V4.1 FlashHigh | 962 | 0.48% | $1,005.8 | ARC-AGI-3 Standard |
| 35 | DeepSeek V4.1 FlashLow | 962 | 0.35% | $820.63 | ARC-AGI-3 Standard |
| 35 | DeepSeek V4.1 FlashNone | 962 | 0.34% | $734.73 | ARC-AGI-3 Standard |
| 35 | Gemini 3.1 Pro (Preview)Not specified | 962 | 0.42% | $2,213.53 | ARC-AGI-3 Standard |
| 35 | GPT-5.5High | 962 | 0.43% | $10,000 | ARC-AGI-3 Standard |
| 35 | GPT-5.6 SolLow | 962 | 0.33% | $12,805.6 | ARC-AGI-3 Standard |
| 35 | GPT-5.6 TerraHigh | 962 | 0.49% | $5,881.39 | ARC-AGI-3 Standard |
| 35 | GPT-6 SolMedium | 962 | 0.42% | $4,863.8 | ARC-AGI-3 Standard |
| 47 | Claude Opus 4.7High | 961 | 0.18% | $10,000 | ARC-AGI-3 Standard |
| 47 | Muse Spark OpenxHigh | 961 | 0.14% | $3,908.42 | ARC-AGI-3 Standard |
| 47 | GPT-5.4High | 961 | 0.21% | $5,187.41 | ARC-AGI-3 Standard |
| 47 | GPT-5.6 LunaHigh | 961 | 0.1% | $2,482.44 | ARC-AGI-3 Standard |
| 47 | GPT-5.6 LunaLow | 961 | 0.17% | $2,307.49 | ARC-AGI-3 Standard |
| 47 | GPT-5.6 LunaMax | 961 | 0.18% | $3,189.42 | ARC-AGI-3 Standard |
| 47 | GPT-5.6 LunaMedium | 961 | 0.17% | $2,379.13 | ARC-AGI-3 Standard |
| 47 | GPT-6 LunaHigh | 961 | 0.18% | $255.98 | ARC-AGI-3 Standard |
| 47 | GPT-6 LunaMax | 961 | 0.1% | $262.24 | ARC-AGI-3 Standard |
| 47 | GPT-6 LunaMedium | 961 | 0.19% | $241.28 | ARC-AGI-3 Standard |
| 47 | GPT-6 LunaxHigh | 961 | 0.16% | $262.68 | ARC-AGI-3 Standard |
| 47 | GPT-6 SolLow | 961 | 0.13% | $4,752 | ARC-AGI-3 Standard |
| 47 | GPT-6 SolNone | 961 | 0.28% | $4,655.55 | ARC-AGI-3 Standard |
| 47 | Grok 4.5High | 961 | 0.3% | $6,892.87 | ARC-AGI-3 Standard |
| 47 | Grok 4.5Low | 961 | 0.26% | $8,718.48 | ARC-AGI-3 Standard |
| 47 | Grok 4.5Medium | 961 | 0.32% | $8,458.3 | ARC-AGI-3 Standard |
| 47 | Grok 4.7Low | 961 | 0.25% | $4,814.88 | ARC-AGI-3 Standard |
| 64 | GPT-5.6 LunaxHigh | 960 | 0.02% | $2,801.59 | ARC-AGI-3 Standard |
| 64 | GPT-5.6 TerraLow | 960 | 0.01% | $5,068.48 | ARC-AGI-3 Standard |
| 64 | GPT-5.6 TerraMedium | 960 | 0.08% | $5,895.31 | ARC-AGI-3 Standard |
| 64 | GPT-6 LunaLow | 960 | 0.03% | $228.83 | ARC-AGI-3 Standard |
| 64 | GPT-6 LunaNone | 960 | 0.03% | $229.77 | ARC-AGI-3 Standard |
| 64 | Grok 4.20 (Beta Reasoning)Not specified | 960 | 0.09% | $3,775 | ARC-AGI-3 Standard |
1 · GPT-6 AstraMax1239
- Performance
- 62.71%
- Evaluation cost · USD
- $26,097.5
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
2 · GPT-6 AstraxHigh1224
- Performance
- 59.34%
- Evaluation cost · USD
- $37,317.39
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
3 · GPT-6 AstraHigh1204
- Performance
- 54.82%
- Evaluation cost · USD
- $40,704.72
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
4 · GPT-6.1 SolMax1195
- Performance
- 52.73%
- Evaluation cost · USD
- $7,632.48
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
5 · Gemini 4 ArgonLow1182
- Performance
- 49.79%
- Evaluation cost · USD
- $5,820.53
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
6 · Gemini 4 ArgonMedium1166
- Performance
- 46.27%
- Evaluation cost · USD
- $6,179.25
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
7 · GPT-6.1 SolxHigh1138
- Performance
- 39.93%
- Evaluation cost · USD
- $9,104.43
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
8 · GPT-6 AstraMedium1132
- Performance
- 38.59%
- Evaluation cost · USD
- $48,090.35
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
9 · GPT-6 AstraNone1117
- Performance
- 35.18%
- Evaluation cost · USD
- $49,791.11
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
10 · Claude Opus 5High1094
- Performance
- 30.16%
- Evaluation cost · USD
- $20,657.37
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
11 · GPT-6.1 SolHigh1079
- Performance
- 26.72%
- Evaluation cost · USD
- $9,757.62
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
12 · Claude Sonnet 5.5xHigh1068
- Performance
- 24.31%
- Evaluation cost · USD
- $7,394.47
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
13 · GPT-6 AstraLow1038
- Performance
- 17.45%
- Evaluation cost · USD
- $38,166.47
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
14 · GPT-6.1 SolMedium1007
- Performance
- 10.57%
- Evaluation cost · USD
- $7,369.78
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
15 · Gemini 3.8 FlashHigh1006
- Performance
- 10.37%
- Evaluation cost · USD
- $4,398.28
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
16 · GPT-5.6 SolMax995
- Performance
- 7.78%
- Evaluation cost · USD
- $25,064.11
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
17 · GPT-5.6 SolxHigh991
- Performance
- 6.99%
- Evaluation cost · USD
- $19,216.38
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
18 · Gemini 3.8 FlashLow987
- Performance
- 5.99%
- Evaluation cost · USD
- $2,669.06
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
19 · GPT-6 SolMax981
- Performance
- 4.62%
- Evaluation cost · USD
- $5,554.49
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
20 · Gemini 3.8 FlashMedium978
- Performance
- 3.99%
- Evaluation cost · USD
- $2,727.66
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
20 · GPT-6.1 SolLow978
- Performance
- 3.92%
- Evaluation cost · USD
- $7,110.56
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
22 · GPT-5.6 SolHigh970
- Performance
- 2.15%
- Evaluation cost · USD
- $15,176.07
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
23 · Grok 4.6xHigh969
- Performance
- 2.11%
- Evaluation cost · USD
- $5,612
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
24 · GPT-6 SolxHigh968
- Performance
- 1.8%
- Evaluation cost · USD
- $5,105.62
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
24 · Grok 4.7High968
- Performance
- 1.73%
- Evaluation cost · USD
- $3,741.65
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
24 · Grok 4.7Medium968
- Performance
- 1.71%
- Evaluation cost · USD
- $3,795.49
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
24 · Grok 4.7xHigh968
- Performance
- 1.82%
- Evaluation cost · USD
- $2,731.58
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
28 · Claude Opus 4.8High967
- Performance
- 1.52%
- Evaluation cost · USD
- $10,000
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
29 · Claude Sonnet 5.5High965
- Performance
- 1.2%
- Evaluation cost · USD
- $5,487.23
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
29 · GPT-5.6 SolMedium965
- Performance
- 1.07%
- Evaluation cost · USD
- $12,971.17
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
31 · DeepSeek V4.1 FlashMax964
- Performance
- 0.85%
- Evaluation cost · USD
- $1,164.91
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
31 · GPT-5.6 TerraMax964
- Performance
- 0.8%
- Evaluation cost · USD
- $7,918.31
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
31 · GPT-6 SolHigh964
- Performance
- 0.9%
- Evaluation cost · USD
- $5,072.45
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
34 · GPT-5.6 TerraxHigh963
- Performance
- 0.65%
- Evaluation cost · USD
- $6,804.04
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
35 · Claude Haiku 5.5Max962
- Performance
- 0.45%
- Evaluation cost · USD
- $388.11
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
35 · Claude Sonnet 5.5Low962
- Performance
- 0.4%
- Evaluation cost · USD
- $4,756.68
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
35 · Claude Sonnet 5.5Medium962
- Performance
- 0.4%
- Evaluation cost · USD
- $4,839.68
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
35 · Claude Opus 4.6Max962
- Performance
- 0.51%
- Evaluation cost · USD
- $8,866.2
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
35 · DeepSeek V4.1 FlashHigh962
- Performance
- 0.48%
- Evaluation cost · USD
- $1,005.8
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
35 · DeepSeek V4.1 FlashLow962
- Performance
- 0.35%
- Evaluation cost · USD
- $820.63
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
35 · DeepSeek V4.1 FlashNone962
- Performance
- 0.34%
- Evaluation cost · USD
- $734.73
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
35 · Gemini 3.1 Pro (Preview)Not specified962
- Performance
- 0.42%
- Evaluation cost · USD
- $2,213.53
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
35 · GPT-5.5High962
- Performance
- 0.43%
- Evaluation cost · USD
- $10,000
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
35 · GPT-5.6 SolLow962
- Performance
- 0.33%
- Evaluation cost · USD
- $12,805.6
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
35 · GPT-5.6 TerraHigh962
- Performance
- 0.49%
- Evaluation cost · USD
- $5,881.39
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
35 · GPT-6 SolMedium962
- Performance
- 0.42%
- Evaluation cost · USD
- $4,863.8
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
47 · Claude Opus 4.7High961
- Performance
- 0.18%
- Evaluation cost · USD
- $10,000
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
47 · Muse Spark OpenxHigh961
- Performance
- 0.14%
- Evaluation cost · USD
- $3,908.42
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
47 · GPT-5.4High961
- Performance
- 0.21%
- Evaluation cost · USD
- $5,187.41
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
47 · GPT-5.6 LunaHigh961
- Performance
- 0.1%
- Evaluation cost · USD
- $2,482.44
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
47 · GPT-5.6 LunaLow961
- Performance
- 0.17%
- Evaluation cost · USD
- $2,307.49
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
47 · GPT-5.6 LunaMax961
- Performance
- 0.18%
- Evaluation cost · USD
- $3,189.42
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
47 · GPT-5.6 LunaMedium961
- Performance
- 0.17%
- Evaluation cost · USD
- $2,379.13
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
47 · GPT-6 LunaHigh961
- Performance
- 0.18%
- Evaluation cost · USD
- $255.98
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
47 · GPT-6 LunaMax961
- Performance
- 0.1%
- Evaluation cost · USD
- $262.24
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
47 · GPT-6 LunaMedium961
- Performance
- 0.19%
- Evaluation cost · USD
- $241.28
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
47 · GPT-6 LunaxHigh961
- Performance
- 0.16%
- Evaluation cost · USD
- $262.68
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
47 · GPT-6 SolLow961
- Performance
- 0.13%
- Evaluation cost · USD
- $4,752
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
47 · GPT-6 SolNone961
- Performance
- 0.28%
- Evaluation cost · USD
- $4,655.55
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
47 · Grok 4.5High961
- Performance
- 0.3%
- Evaluation cost · USD
- $6,892.87
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
47 · Grok 4.5Low961
- Performance
- 0.26%
- Evaluation cost · USD
- $8,718.48
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
47 · Grok 4.5Medium961
- Performance
- 0.32%
- Evaluation cost · USD
- $8,458.3
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
47 · Grok 4.7Low961
- Performance
- 0.25%
- Evaluation cost · USD
- $4,814.88
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
64 · GPT-5.6 LunaxHigh960
- Performance
- 0.02%
- Evaluation cost · USD
- $2,801.59
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
64 · GPT-5.6 TerraLow960
- Performance
- 0.01%
- Evaluation cost · USD
- $5,068.48
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
64 · GPT-5.6 TerraMedium960
- Performance
- 0.08%
- Evaluation cost · USD
- $5,895.31
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
64 · GPT-6 LunaLow960
- Performance
- 0.03%
- Evaluation cost · USD
- $228.83
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
64 · GPT-6 LunaNone960
- Performance
- 0.03%
- Evaluation cost · USD
- $229.77
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
64 · Grok 4.20 (Beta Reasoning)Not specified960
- Performance
- 0.09%
- Evaluation cost · USD
- $3,775
- Agent / harness
- ARC-AGI-3 Standard
- Evidence
- Model profile ↗
Source & conversion
Semi-Private evaluations. Standard is the primary harness; Provider Adapter results remain separate and never substitute for Standard.
The primary score measures human-normalized skill-acquisition performance, not accuracy. Costs are published evaluation totals, not per-task prices. Notes retain any partial-usage limitation.
How the Lumina conversion works
1,000 + 75 × ((original score − 8.97312670612202) / 16.8510495584195). The result is rounded once. Scores may tie.
Frozen initial verified cohort of 69 configurations; Standard harness only. Provider Adapter uses the same fixed scale but stays a separate view.
- Calibration
- arc-agi-3-z75-v1 · experimental
- Reference mean
- 8.97312670612202
- Population SD
- 16.8510495584195
- Coverage
- Standalone Reasoning benchmark.
Identity, exclusions & attribution
Verified publisher data, not an independent Lumina rerun.
Official leaderboard · Benchmark description
Publisher terms. Scores and resource figures are attributed to the original publisher.