SREBench leaderboard
As of 6 Nov 2026, the top score on SREBench is 56.9% by GPT-6 Astra (OpenAI), ahead of GPT-6.1 Sol at 50.8%, across 37 model configurations with a published result (Vals AI, result dated 5 Nov 2026). An agent is given a broken system and must find and fix the fault, in the style of on-call engineering. SREBench is shown for reference only and does not count toward the BenchLeader Index.
Key facts (as of 6 Nov 2026)
- GPT-6 Astra holds the top SREBench score at 56.9% (Vals AI, 5 Nov 2026).
- GPT-6.1 Sol is second on SREBench at 50.8% (Vals AI, 5 Nov 2026).
- SREBench: Site-reliability incidents to diagnose and fix. Run by Vals AI (Vals AI).
- 37 model configurations have a published SREBench result on BenchLeader (BenchLeader, 6 Nov 2026).
- SREBench does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).
Site-reliability incidents to diagnose and fix. Run by Vals AI.
- Published by
- Vals AI
- Category
- Agents & tools
- Index weight
- Reference only
- Models
- 37
- Data as of
- 6 Nov 2026
Vals AI (vals.ai).
What the test looks like
An agent is given a broken system and must find and fix the fault, in the style of on-call engineering.
How it is scored
Percent of incidents resolved, run by Vals AI.
What to keep in mind
Few models measured; long tasks make scores volatile.
- 1GPT-6 Astra (max)56.9%
- 2GPT-6.1 Sol (max)50.8%
- 3Gemini 4 Argon (high)44.3%
- 4Claude Opus 5.533.6%
- 5Daybreak Blue (max)31.7%
- 6GPT-5.6 Sol (max)30.5%
- 7GPT-6 Sol (max)30.5%
- 8Claude Sonnet 5.530.1%
- 9Kimi K3.1 Pro (max)27.1%
- 10Claude Fable 5.122.9%
- 11Claude Opus 5.5 Range21.8%
- 12Claude Opus 512.2%
- 13Muse Wonder Full (max)11.8%
- 14Gemini 3.8 Flash (high)11.1%
- 15GPT-5.6 Terra (max)9.5%
37 of 37
| # | |||||
|---|---|---|---|---|---|
| 1 |  GPT-6 Astramax | 56.9% | 67.8 | $10.24 | 2026-11-05 |
| 2 | GPT-6.1 Solmax | 50.8% | 67.3 | $2.69 | 2026-11-05 |
| 3 |  Gemini 4 Argonhigh | 44.3% | 66.0 | $30.45 | 2026-11-05 |
| 4 |  Claude Opus 5.5 | 33.6% | 63.4 | $34.65 | 2026-11-05 |
| 5 | Daybreak Bluemax | 31.7% | – | $54.34 | 2026-11-05 |
| 6 | GPT-5.6 Solmax | 30.5% | 65.4 | $42.52 | 2026-11-05 |
| 7 | GPT-6 Solmax | 30.5% | 63.3 | $20.65 | 2026-11-05 |
| 8 | Claude Sonnet 5.5 | 30.1% | – | $26.52 | 2026-11-05 |
| 9 |  Kimi K3.1 Promaxopen ↗ | 27.1% | – | $27.39 | 2026-11-05 |
| 10 | Claude Fable 5.1 | 22.9% | 66.4 | $32.23 | 2026-11-05 |
| 11 | Claude Opus 5.5 Range | 21.8% | – | $33.13 | 2026-11-05 |
| 12 | Claude Opus 5 | 12.2% | 65.2 | $23.63 | 2026-11-05 |
| 13 |  Muse Wonder Fullmax | 11.8% | 62.9 | $33.81 | 2026-11-05 |
| 14 | Gemini 3.8 Flashhigh | 11.1% | 62.0 | $27.46 | 2026-11-05 |
| 15 | GPT-5.6 Terramax | 9.5% | 60.7 | $30.86 | 2026-11-05 |
| 16 |  GLM 5.3maxopen ↗ | 9.2% | 59.0 | $27.47 | 2026-11-05 |
| 17 | Claude Haiku 5.5 | 5.7% | – | $3.14 | 2026-11-05 |
| 18 | Gemini 3.7 Flashhigh | 4.6% | 62.2 | $21.47 | 2026-11-05 |
| 19 |  MiMo-V2.6-Flashopen ↗ | 4.2% | 55.0 | $0.395 | 2026-11-05 |
| 20 | GPT-5.5xhigh | 3.8% | 64.3 | $17.77 | 2026-11-05 |
| 21 | Muse Spark 1.3max | 3.8% | 62.0 | $25.86 | 2026-11-05 |
| 22 |  Grok 4.7xhigh | 3.8% | 58.3 | $64.77 | 2026-11-05 |
| 23 | MiMo-V2.6-Proopen ↗ | 3.0% | 57.3 | $0.533 | 2026-11-05 |
| 24 | GPT-6 Lunamax | 2.7% | 56.7 | $1.55 | 2026-11-05 |
| 25 |  Hy4open ↗ | 2.3% | 55.7 | $4.90 | 2026-11-05 |
| 26 | NewInkling 2 Small | 1.9% | 57.4 | $6.67 | 2026-11-05 |
| 27 |  DeepSeek V4 Promaxopen ↗ | 1.9% | 57.0 | $5.58 | 2026-11-05 |
| 28 | GPT-5.6 Lunamax | 1.5% | 56.3 | $3.44 | 2026-11-05 |
| 29 | DeepSeek V4.1 Flashmaxopen ↗ | 0.8% | 59.0 | $0.549 | 2026-11-05 |
| 30 | Grok 4.5high | 0.8% | 56.9 | $13.42 | 2026-11-05 |
| 31 | GLM 5.3 Flashmaxopen ↗ | 0.8% | 54.2 | $2.29 | 2026-11-05 |
| 32 | Claude Sonnet 5 | 0.4% | 55.9 | $18.49 | 2026-11-05 |
| 33 |  Step 5 Previewopen | 0.4% | 54.9 | $3.07 | 2026-11-05 |
| 34 | Muse Spark 1.2xhigh | 0.0% | 61.7 | $23.43 | 2026-11-05 |
| 35 | Gemini 3.1 Prohigh | 0.0% | 60.1 | $11.34 | 2026-11-05 |
| 36 | GLM 5.2maxopen ↗ | 0.0% | 59.5 | $26.58 | 2026-11-05 |
| 37 | Inklingopen ↗ | 0.0% | 47.5 | $2.23 | 2026-11-05 |
More agents & tools benchmarks · All LLM benchmarks
Related benchmarks
Top head-to-heads on SREBench
Cite as: BenchLeader, “SREBench leaderboard”, data as of 6 Nov 2026, https://www.benchleader.com/benchmarks/vals_srebench.
SREBench: questions
- What does SREBench measure?
- An agent is given a broken system and must find and fix the fault, in the style of on-call engineering. Scores are reported in percent of tasks solved; higher is better.
- What is the top score on SREBench?
- As of 6 Nov 2026, the highest score on SREBench is 56.9% by GPT-6 Astra (result dated 5 Nov 2026), ahead of GPT-6.1 Sol at 50.8%. Scores come from Vals AI and are re-read daily.
- Which AI model leads SREBench?
- GPT-6 Astra leads SREBench with 56.9% as of 6 Nov 2026 (result dated 5 Nov 2026), ahead of GPT-6.1 Sol at 50.8%.
- How many models have SREBench results?
- 37 model configurations have a SREBench result on BenchLeader, all taken from Vals AI.
- Who runs SREBench and how often is it updated?
- SREBench is published by Vals AI. BenchLeader re-reads the published results once a day and records the date each result was published.
- Does SREBench count toward the BenchLeader Index?
- No. SREBench is shown for reference but left out of the composite index.