Scale Labs SWE Atlas - Refactoring (SWE Atlas - Refactoring)
We show this table for reference; we do not rank on it.
A Scale Labs public leaderboard mirrored as display-only reference data. It does not affect BenchLM rankings.
Scale score on SWE Atlas - Refactoring — November 12, 2026 snapshot
We mirror the published scale score view for SWE Atlas - Refactoring. Fable 5.2 leads the public snapshot at 80%, followed by Opus 5.5 (77.1%) and Fable 5.1 (71.4%). We do not use these results to rank models overall.
Fable 5.2
Anthropic
SWE Atlas - Refactoring
Opus 5.5
Anthropic
SWE Atlas - Refactoring
Fable 5.1
Anthropic
SWE Atlas - Refactoring
28 modelsCodingCurrentDisplay onlyUpdated November 12, 2026 snapshot
Scale score table (28 models)
| Rank | Model / configuration | Score | Parameters (B) | Open / closed |
|---|---|---|---|---|
| 1 | Fable 5.2AnthropicSWE Atlas - Refactoring | 80% | Not reported | Unknown |
| 2 | Opus 5.5AnthropicSWE Atlas - Refactoring | 77.1% | Not reported | Unknown |
| 3 | Fable 5.1AnthropicSWE Atlas - Refactoring | 71.4% | Not reported | Unknown |
| 4 | GPT-6 AstraOpenAISWE Atlas - Refactoring | 71.4% | Not reported | Closed |
| 5 | Gemini 4 ArgonGoogleSWE Atlas - Refactoring | 67.1% | Not reported | Closed |
| 6 | Gemini 4 CarbonGoogleSWE Atlas - Refactoring | 64.3% | Not reported | Closed |
| 7 | Sonnet 5.5AnthropicSWE Atlas - Refactoring | 62.9% | Not reported | Unknown |
| 8 | GPT-6.1 SolOpenAISWE Atlas - Refactoring | 60% | Not reported | Closed |
| 9 | Muse Wonder 1.0MetaSWE Atlas - Refactoring | 60% | Not reported | Unknown |
| 10 | Gemini 3.8 FlashGoogleSWE Atlas - Refactoring | 55.7% | Not reported | Closed |
| 11 | GPT-6.1 TerraOpenAISWE Atlas - Refactoring | 55.7% | Not reported | Closed |
| 12 | DeepSeek V4.1 Flash 1017DeepSeekSWE Atlas - Refactoring | 54.3% | Not reported | Open |
| 13 | Kimi K3Moonshot AISWE Atlas - Refactoring | 54.3% | Not reported | Open |
| 14 | Kimi K3.1 ProMoonshot AISWE Atlas - Refactoring | 54.3% | Not reported | Open |
| 15 | Inkling 2 SmallThinking Machines LabSWE Atlas - Refactoring | 53.3% | Not reported | Closed |
| 16 | Haiku 5.5AnthropicSWE Atlas - Refactoring | 52.9% | Not reported | Unknown |
| 17 | Qwen 3.8 MaxAlibabaSWE Atlas - Refactoring | 52.9% | Not reported | Unknown |
| 18 | Muse Ember 1.0MetaSWE Atlas - Refactoring | 50% | Not reported | Unknown |
| 19 | GLM-5.3Z.AISWE Atlas - Refactoring | 48.6% | Not reported | Open |
| 20 | Grok 4.7xAISWE Atlas - Refactoring | 48.6% | Not reported | Closed |
| 21 | GPT-6 LunaOpenAISWE Atlas - Refactoring | 47.1% | Not reported | Closed |
| 22 | GLM-5.3-FlashZ.AISWE Atlas - Refactoring | 45.7% | Not reported | Open |
| 23 | Muse Spark OpenMetaSWE Atlas - Refactoring | 35.7% | Not reported | Open |
| 24 | Mistral Large 4MistralSWE Atlas - Refactoring | 34.3% | Not reported | Open |
| 25 | MiniMax M3MiniMaxSWE Atlas - Refactoring | 28.6% | Not reported | Open |
| 26 | Nemotron 3.5 Super VLNvidiaSWE Atlas - Refactoring | 20% | Not reported | Unknown |
| 27 | Nemotron 3 UltraNVIDIASWE Atlas - Refactoring | 15.7% | Not reported | Open |
| 28 | GPT-OSS 120BOpenAISWE Atlas - Refactoring | 1.4% | Not reported | Open |
How to read this leaderboard
Compare the published configurations as complete evaluation systems. The source can combine a base model, agent scaffold, tools, budget, and inference setting in each result.
Operator receipt: 28 sourced rows are currently displayable on this page; the leading published row is Fable 5.2 at 80%.
Honest limit: This Scale table is display-only context, not benchmark provenance or a weighted model-only comparison.
How BenchLM shows SWE Atlas - Refactoring
BenchLM mirrors 28 published rows from Scale Labs’ public SWE Atlas - Refactoring leaderboard, captured on November 12, 2026 snapshot.
The table is display only. It is useful context for a published agent or model configuration, but it does not enter BenchLM’s overall or category rankings.
Snapshot
The published SWE Atlas - Refactoring snapshot places Fable 5.2 first at 80%. The third row is 8.6 points behind. The broader top-10 range is 24.3 points, so the table still separates the published systems.
28 models have been evaluated on SWE Atlas - Refactoring. The benchmark falls in the Coding category. We keep external benchmark mirrors separate from the weighted global scoring system, so these results remain source-specific evidence. SWE Atlas - Refactoring is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
About SWE Atlas - Refactoring
Year
2026
Tasks
28 published rows
Format
Published Scale leaderboard score
Difficulty
External agent and model evaluation
BenchLM mirrors 28 published rows from the SWE Atlas - Refactoring public table captured on November 12, 2026 snapshot.
Freshness and provenance
Version
SWE Atlas - Refactoring 2026
Refresh cadence
Quarterly
Staleness state
Current
Question availability
Public benchmark set
BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.
Questions
What does SWE Atlas - Refactoring measure?
A Scale Labs public leaderboard mirrored as display-only reference data. It does not affect BenchLM rankings.
Which model leads the published SWE Atlas - Refactoring snapshot?
Fable 5.2 currently leads the published SWE Atlas - Refactoring snapshot with 80% scale score. BenchLM shows this benchmark for display only and does not use it in overall rankings.
How many models are evaluated on SWE Atlas - Refactoring?
The November 12, 2026 snapshot snapshot contains 28 AI models.
Know when it’s worth switching models
The model to choose, the cheaper alternative, and the release we would wait on.
Read a sample issueJoin 7,000+ subscribers.
One email each week. Unsubscribe anytime.