- Context
- 1Mtokens
- Released
- 5mo ago
Cost/1M tokens
Benchmarks
DeepSWE is the primary coding benchmark where an exact model evaluation exists. Arena leaderboards and other sourced evaluations appear below; missing scores are not zero.
How it ranks
Against its predecessor, similarly priced models and the models ranked around it.
DeepSWE 1.1
Best pass@1 per model · Higher is better
#5 of 28
- 74.1%
- 73.8%
- 73.6%
- 72.7%
- 69.9%Claude Fable 5
- 69.6%
- 69.0%
- 68.5%
- 67.5%
- 67.2%
Arena Text
Best Arena rating per model · Higher is better
Arena #5
- 1511
Muse WonderNearby rank - 1506Claude Opus 5.5Nearby rank
- 1504Claude Fable 5
- 1503Claude Opus 4.6Nearby rank
- 1503Claude Fable 5.1Similar price
- 1481GPT-5.5Similar price
- 1474GPT-6 AstraSimilar price
- 1450Claude Opus 4.1Similar price
Ahead of 3 of 4 similarly priced models
Arena WebDev
Best Arena rating per model · Higher is better
Arena #26
- 1836Claude Fable 5.2Similar price
- 1764GPT-6 AstraSimilar price
- 1742Claude Fable 5.1Similar price
- 1638
MiMo-V2.6-FlashNearby rank - 1632
Hy4 previewNearby rank - 1625Claude Fable 5
- 1624MiMo-V2.6-ProNearby rank
- 1624
Qwen3.8 Flash NextNearby rank - 1512GPT-5.5Similar price
Ahead of 1 of 4 similarly priced models
Arena Vision
Best Arena rating per model · Higher is better
Arena #1
- 1309Claude Fable 5
- 1300Claude Opus 4.7Similar price
- 1299Qwen3.8 MaxNearby rank
- 1289Claude Fable 5.1Similar price
- 1285GPT-5.5Similar price
- 1283GPT-6 AstraSimilar price
Ahead of 3 of 4 similarly priced models
Arena Search
Best Arena rating per model · Higher is better
Arena #5
- 1245GPT-6 AstraSimilar price
- 1229GPT-6 SolNearby rank
- 1229Claude Fable 5
- 1227Qwen3.8 MaxNearby rank
- 1227Claude Fable 5.1Similar price
- 1213Claude Opus 5Similar price
- 1210Claude Opus 4.7Similar price
Ahead of 2 of 4 similarly priced models
Arena Document
Best Arena rating per model · Higher is better
Arena #1
- 1500Claude Fable 5
- 1498Claude Fable 5.1Similar price
- 1497Claude Opus 4.6Similar price
- 1481GPT-5.5Similar price
- 1474GPT-6 AstraSimilar price
Ahead of 2 of 4 similarly priced models
Agent Arena
Best IPS score per model · Higher is better
Arena #11
- +0.304Claude Fable 5.2Similar price
- +0.148Claude Fable 5.1Similar price
- +0.145GPT-6 AstraSimilar price
- +0.107Gemini 4 ArgonNearby rank
- +0.090Muse Wonder FullNearby rank
- +0.073Claude Fable 5
- +0.072Claude Opus 5Similar price
- +0.069GPT-6 SolNearby rank
Ahead of 0 of 4 similarly priced models
ARC-AGI-2
Best score per model · Higher is better
#10 of 85
- 95.0%GPT-6 AstraSimilar price
- 90.0%Claude Fable 5.1Similar price
- 89.6%GPT-6 SolNearby rank
- 89.2%Claude Fable 5
- 89.2%Gemini 3.8 FlashNearby rank
- 88.3%Claude Opus 5Nearby rank
- 85.0%GPT-5.5Similar price
- 4.9%o3 ProSimilar price
Ahead of 2 of 4 similarly priced models
FrontierMath-Tier-4-v2-Private
Best score per model · Higher is better
#7 of 72
- 100.0%Claude Fable 5.2Similar price
- 97.6%GPT-6 AstraSimilar price
- 95.0%Claude Opus 5.5Nearby rank
- 92.7%GPT-6.1 TerraNearby rank
- 90.2%Claude Fable 5
- 90.0%GPT-6 SolNearby rank
- 87.8%Claude Fable 5.1Similar price
- 72.5%GPT-5.5Similar price
Ahead of 2 of 4 similarly priced models
FrontierMath-Tiers-1-3-v2-Private
Best score per model · Higher is better
#12 of 87
- 95.1%Claude Fable 5.2Similar price
- 93.7%GPT-6 AstraSimilar price
- 90.2%Claude Fable 5.1Similar price
- 88.8%Claude Sonnet 5.5Nearby rank
- 87.7%GPT-5.5 ProNearby rank
- 87.0%Claude Fable 5
- 86.0%Gemini 4 ArgonNearby rank
- 86.0%GPT-5.6 TerraNearby rank
- 85.3%GPT-5.5Similar price
Ahead of 1 of 4 similarly priced models
GPQA diamond
Best score per model · Higher is better
#67 of 161
- 95.8%GPT-6 AstraSimilar price
- 93.9%Claude Fable 5.2Similar price
- 90.7%GPT-5.5Similar price
- 86.2%GPT-5Nearby rank
- 86.0%Claude Opus 4.5Nearby rank
- 85.9%Claude Fable 5
- 85.9%
Qwen3.6 27BNearby rank - 85.4%
Nemotron 3 UltraNearby rank - 77.3%Claude Opus 4.1Similar price
Ahead of 1 of 4 similarly priced models
OTIS Mock AIME 2024-2025
Best score per model · Higher is better
#1 of 158
- 100.0%Claude Fable 5
- 100.0%Claude Fable 5.1Similar price
- 100.0%Claude Opus 5.5Nearby rank
- 100.0%GPT-6 AstraSimilar price
- 99.7%Claude Fable 5.2Similar price
- 84.4%GPT-5.5Similar price
Ahead of 2 of 4 similarly priced models
SimpleQA Verified
Best score per model · Higher is better
#9 of 87
- 94.0%Claude Fable 5.2Similar price
- 75.6%GPT-6 AstraSimilar price
- 72.2%Claude Opus 5.5Nearby rank
- 70.8%Claude Fable 5.1Similar price
- 70.7%Claude Fable 5
- 69.7%Gemini 3.8 FlashNearby rank
- 69.7%GPT-5.6 SolNearby rank
- 63.0%GPT-5.5Similar price
Ahead of 1 of 4 similarly priced models
DeepSWE 1.1
Primary coding-agent benchmark · 113 tasks · Data generated 2026-09-22
- pass@4
- 88.5%
- Average cost / attempt
- $13.41
- Benchmark runs
- 4
Score vs cost by reasoning effort
Average cost per attempt (USD) against pass rate
All published configurations
| Harness / reasoning | pass@1 | pass@4 | Cost / attempt |
|---|---|---|---|
| mini-swe-agent · xhigh | 69.9% | 88.5% | $13.41 |
| mini-swe-agent · max | 69.7% | 84.1% | $21.63 |
| mini-swe-agent · high | 68.6% | 86.7% | $9.18 |
| mini-swe-agent · medium | 65.4% | 83.2% | $6.09 |
| mini-swe-agent · low | 59.6% | 81.4% | $3.76 |
pass@1 is the success rate across scored attempts; pass@4 is the share of tasks solved in at least one of four attempts. Costs are reported by DeepSWE for each configuration. These results are not directly comparable to evaluations using different harnesses. Official source JSON.
Reasoning effort
ARC-AGI-2
Score by reasoning effort · Higher is better
GPQA diamond
Score by reasoning effort · Higher is better
OTIS Mock AIME 2024-2025
Score by reasoning effort · Higher is better
All published scores
- Arena TextArena #51504
- high150495% CI 1500 to 1508 · 45,524 votes
- Arena WebDevArena #261625
- high162595% CI 1619 to 1631 · 16,998 votes
- Arena VisionArena #11309
- high130995% CI 1303 to 1316 · 15,536 votes
- Arena SearchArena #51229
- high122995% CI 1222 to 1235 · 69,528 votes
- Arena DocumentArena #11500
- high150095% CI 1493 to 1507 · 12,357 votes
- Agent ArenaArena #11+0.073
- high+0.07395% CI +0.062 to +0.085 · 3,549,956 observations
- ARC-AGI-2#10 of 8589.2%
- low76.8%
- medium82.5%
- high87.5%
- xhigh88.3%
- max89.2%
- FrontierMath-Tier-4-v2-Private#7 of 7290.2%
- max90.2%
Source:Epoch AI evaluation - FrontierMath-Tiers-1-3-v2-Private#12 of 8787.0%
- max87.0%
Source:Epoch AI evaluation - GPQA diamond#67 of 16185.9%
- low78.8%
- high83.3%
- max85.9%
Source:Epoch AI evaluation - OTIS Mock AIME 2024-2025#1 of 158100.0%
- low97.8%
- high100.0%
- max99.7%
Source:Epoch AI evaluation - SimpleQA Verified#9 of 8770.7%
- xhigh70.7%
Source:Epoch AI evaluation
Benchmark compilation: Epoch AI, Capabilities & benchmarking (CC BY attribution). Scores are converted to percentages and model names matched to the Vertical Tabs catalog. Individual evaluations are credited to their original sources. External results may carry separate licences and evaluation configurations; do not interpret scores from different harnesses as directly comparable.
Arena ratings: Arena (lmarena-ai/leaderboard-dataset), licensed CC BY 4.0. Ratings come from pairwise human votes and are only comparable within one leaderboard; the interval is Arena's 95% confidence interval. Model names are matched to the Vertical Tabs catalog.