Benchmark Almanac

Incident Arena

GPT-6 Astra 64.3% leads since September 2026
Not in model cards

Agents diagnose and repair incidents in running services. The score averages the repair success rate across 20 incidents.

Released
September 2026
Kind
Terminal tasks
Level
Expert
Scored
Checked automatically
Answers
Agent task
Tasks
All public
Measures
Capability
Language
English
In and out
Text, Code → Actions, Code
Interaction
Agentic
Environment
Terminal
Refresh
One fixed set
Run by
Anyone
Occupations
Computer and mathematical
In views
Agents
Strongest model tested
GPT-6 Astra #4 on the capability index
Last model added
October 2026
Scores refreshed
November 20th, 2026
15
Entries on its own leaderboard, 14 models we track

GPT-6 Astra leads with 64.3% GPT-6 Astra leads with 64.3%

Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.

  1. 1 GPT-6 Astra xhigh 64.3%
  2. 2 Claude Opus 5.5 max 63.6%
  3. 3 GPT-6.1 Sol medium 50.7%
  4. 4 Gemini Carbon max 45.7%
  5. 5 GPT-5.6 Terra max 45.7%
  6. 6 GPT-6 Sol medium 44.4%
  7. 7 Claude Fable 5.1 max 43.8%
  8. 8 Muse Wonder xhigh 40%
  9. 9 Claude Opus 5 low 38.6%
  10. 10 GPT-5.6 Sol high 37.1%
  11. 11 Muse Spark 1.3 max 36.4%
  12. 12 Claude Sonnet 5.5 high 31%
  13. 13 GLM-5.3 high 30%
  14. 14 SWE-2 27.1%
  15. 15 Grok 4.7 medium 23.6%

Source: Official Incident Arena leaderboard

#EntryScore ±SettingDateLinkSource
1 GPT-6 Astra 64.3% 45 to 82.1 xhigh Sep 2026 Official leaderboard
2 Claude Opus 5.5 63.6% 47.9 to 78.6 max Sep 2026 Official leaderboard
3 GPT-6.1 Sol 50.7% ±17.1 medium Sep 2026 Official leaderboard
4 Gemini Carbon 45.7% ±17.1 max oddish.app Official leaderboard
5 GPT-5.6 Terra 45.7% 30.7 to 61.4 max Jul 2026 Official leaderboard
6 GPT-6 Sol 44.4% ±16.8 medium Sep 2026 Official leaderboard
7 Claude Fable 5.1 43.8% ±13.9 max Sep 2026 Official leaderboard
8 Muse Wonder 40% 22.9 to 58.6 xhigh Oct 2026 Official leaderboard
9 Claude Opus 5 38.6% ±15 low Jul 2026 Official leaderboard
10 GPT-5.6 Sol 37.1% 22.9 to 52.9 high Jul 2026 Official leaderboard
11 Muse Spark 1.3 36.4% 20 to 53.6 max Sep 2026 Official leaderboard
12 Claude Sonnet 5.5 31% 19.1 to 43.8 high Sep 2026 Official leaderboard
13 GLM-5.3 30% 17.1 to 44.3 high Aug 2026 Official leaderboard
14 SWE-2 27.1% 15 to 40 Sep 2026 oddish.app Official leaderboard
15 Grok 4.7 23.6% 11.4 to 37.1 medium Sep 2026 Official leaderboard
#ModelScore ±SettingDateLinkSource
1 GPT-6 Astra 64.3% 45 to 82.1 xhigh Sep 2026 Official leaderboard
2 Claude Opus 5.5 63.6% 47.9 to 78.6 max Sep 2026 Official leaderboard
3 GPT-6.1 Sol 50.7% ±17.1 medium Sep 2026 Official leaderboard
4 GPT-5.6 Terra 45.7% 30.7 to 61.4 max Jul 2026 Official leaderboard
5 GPT-6 Sol 44.4% ±16.8 medium Sep 2026 Official leaderboard
6 Claude Fable 5.1 43.8% ±13.9 max Sep 2026 Official leaderboard
7 Muse Wonder 40% 22.9 to 58.6 xhigh Oct 2026 Official leaderboard
8 Claude Opus 5 38.6% ±15 low Jul 2026 Official leaderboard
9 GPT-5.6 Sol 37.1% 22.9 to 52.9 high Jul 2026 Official leaderboard
10 Muse Spark 1.3 36.4% 20 to 53.6 max Sep 2026 Official leaderboard
11 Claude Sonnet 5.5 31% 19.1 to 43.8 high Sep 2026 Official leaderboard
12 GLM-5.3 30% 17.1 to 44.3 high Aug 2026 Official leaderboard
13 SWE-2 27.1% 15 to 40 Sep 2026 oddish.app Official leaderboard
14 Grok 4.7 23.6% 11.4 to 37.1 medium Sep 2026 Official leaderboard

Rankings by part

We read 20 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as 000--frappe--07-deletes-and-jobs-fail-194d9279, 001--frappe--07-desk-and-queue-oom-0669b8e8 and 002--frappe--07-desk-and-queue-outage-64526986. GPT-6 Astra leads 6 of them, and 7 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.

000--frappe--07-deletes-and-jobs-fail-194d9279

14 models we track have a score here.

  1. 1 GPT-5.6 Terra ↑3 100%
  2. 2 Claude Opus 5 ↑6 100%
  3. 3 GPT-6 Astra ↓2 100%
  4. 4 GPT-6.1 Sol ↓1 100%
  5. 5 Muse Wonder ↑2 100%
  6. 6 Muse Spark 1.3 ↑4 85.7%
  7. 7 Grok 4.7 ↑7 71.4%
  8. 8 GLM-5.3 ↑4 57.1%
  9. 9 Claude Opus 5.5 ↓7 57.1%
  10. 10 GPT-6 Sol ↓5 57.1%

2 models have held number one on the official leaderboard since its release in September 2026

Lab claims count from the day the card came out, independent runs from the model’s release.

OpenAI
Own leaderboard 2 leaders
GPT-6 Astra 64.3%

Source: Official Incident Arena leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
On the benchmark’s own leaderboard Best so far
0% 20% 40% 60% 80% 100% Near the ceiling (85%) GPT-6 Astra GPT-5.6 Terra 0% 20% 40% 60% 80% 100% Near the ceiling (85%) GPT-6 Astra GPT-5.6 Terra

Source: Official Incident Arena leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

More in Terminal agents

All 35 in Terminal agents
  • SkillsBench In use 86-87 containerized expert tasks in 11 domains, run with and without Agent Skills. Verifier tests give pass rate and skill lift. GPT-5.5 66.5%
  • Terminal-Bench Hard In use Agents complete 44 hard tasks from Terminal-Bench Core in the Terminus 2 harness. Artificial Analysis scores pass@1 averaged over 3 repeats. Gemma 4 31B IT 36%
  • Terminal-Bench 4.1 In use The Terminal-Bench 4.0 tasks with hardened verifiers: agents complete 66 hard tasks in containerized terminals across software, ML, science, security, and operations. Test scripts verify each result, scored as resolution rate. Copilot Fusion 71.2%
  • Harbor-Index In use Agents solve about 80 hard tasks drawn from 29 benchmarks run through Harbor adapters, such as HLE, GSO and SWE-bench Pro. Each task keeps its source verifier. GPT-5.5 28.1%
  • Terminal-Bench 2.1 Extended Not in model cards Agents complete long command-line tasks in containers, scored by verification tests. Mercor adds 99 new expert-written tasks to the 89 public ones and reports both scores. Muse Wonder 54%
  • Terminal-Bench Pro Not in model cards Agents solve 400 terminal tasks across eight domains. Harbor distributes the 200 public tasks. Alibaba keeps the other 200 for private evaluation. Claude Opus 4.5 51%

Data audit

All 11 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 5 pass
  • Pass Independent runs Scores from the official leaderboard (15 entries).
  • Pass Models scored 14 models we track, plus 1 other entry on its leaderboard.
  • Pass Newest score The newest is from October 2026, for Muse Wonder.
  • Pass Frontier models 5 of the top 10 have a score, led by GPT-6 Astra (#4).
  • Pass Headroom The best is 64.3%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on incident-arena.com: 14 rows for models we track and 1 other entry at the last read.
  • Pass Last read November 20th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check