Benchmark Almanac

EnterpriseOps-Gym

Claude Opus 4.6 45.9% leads since March 2026
Inkling 2 Small 46.8% leads since November 2026
In use
Version Original

The official leaderboard hasn’t scored a model released after Gemini 3.1 Pro (February 2026). Labs report their own numbers for 4 newer models, such as Inkling 2 Small at 46.8%.

Agents do 1,150 expert-written tasks in eight enterprise domains, such as IT, HR and customer service, using 512 tools over live MCP servers. SQL checks on the final database state score each task.

Some tasks are impossible and test refusal. Artificial Analysis runs its own version named EnterpriseOps-Gym-AA.

Made by
ServiceNow
Released
March 2026
Kind
Apps and MCP servers
Level
Expert
Scored
Checked automatically
Answers
Agent task
Tasks
All public
Measures
Capability
Language
English
In and out
Text → Text, Actions
Interaction
Agentic
Environment
APIs and tools
Items from
Written by experts
Run by
Anyone
Strongest model tested
Claude Opus 4.6 #39 on the capability index
Last in a launch post
November 2026
Last model added
November 2026
Scores refreshed
November 18th, 2026

In the authors’ words

Large language models are shifting from passive information providers to active agents intended for complex workflows. However, their deployment as reliable AI workers in enterprise is stalled by benchmarks that fail to capture the intricacies of professional environments, specifically, the need for long-horizon planning amidst persistent state changes and strict access protocols. In this work, we introduce EnterpriseOps-Gym, a benchmark designed to evaluate agentic planning in realistic enterprise settings. Specifically, EnterpriseOps-Gym features a containerized sandbox with 164 database tables and 512 functional tools to mimic real-world search friction. Within this environment, agents are evaluated on 1,150 expert-curated tasks across eight mission-critical verticals (including Customer Service, HR, and IT). Our evaluation of 14 frontier models reveals critical limitations in state-of-the-art models: the top-performing Claude Opus 4.5 achieves only 37.4% success. Further analysis shows that providing oracle human plans improves performance by 14-35 percentage points, pinpointing strategic reasoning as the primary bottleneck. Additionally, agents frequently fail to refuse infeasible tasks (best model achieves 53.9%), leading to unintended and potentially harmful side effects. Our findings underscore that current agents are not yet ready for autonomous enterprise deployment. More broadly, EnterpriseOps-Gym provides a concrete testbed to advance the robustness of agentic planning in professional workflows.

Abstract of EnterpriseOps-Gym: Environments and Evaluations for Stateful Agentic Planning and Tool Use in Enterprise Settings (arXiv:2603.13594, CC0)
15
Entries on its own leaderboard, 12 models we track
4
Launch posts list it, from 3 labs
1
System card reports it, from 1 lab

Claude Opus 4.6 leads with 45.9% among models released by February 2026 Claude Opus 4.6 leads with 45.9% among models released by February 2026

Best score per entry from the benchmark's own leaderboard, read on November 18th, 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.

Inkling 2 Small leads with 46.8%, claims included

Best score per model from every source, including labs’ own numbers and claims Epoch AI collected from reports and papers. Labs pick their own settings, so a claim can beat an independent run of the same model. Striped bars are claims, and dotted bars are reruns by Epoch AI or another board.

The newest model with a score here came out in February 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 Claude Opus 4.6 Oracle 45.9%
  2. 2 Claude Sonnet 4.6 Oracle 42.2%
  3. 3 Claude Opus 4.5 Oracle 39.4%
  4. 4 Gemini 3.1 Pro Oracle 38.7%
  5. 5 Claude Sonnet 4.5 Oracle 34.1%
  6. 6 Gemini 3 Flash Oracle 34%
  7. 7 GPT-5 Oracle 30.9%
  8. 8 Gemini 3 Pro Oracle 29.7%
  9. 9 DeepSeek-V3.2 Oracle 24.2%
  10. 10 GPT-5 mini Oracle 22.9%
  11. 11 Kimi K2 Thinking Oracle 22.4%
  12. 12 Gemini-2.5-Pro (snapshot unspecified) Oracle 20%
  13. 13 Qwen3-30B Oracle; Think 19.1%
  14. 14 Qwen3-235B Oracle; Inst. 18.1%
  15. 15 Qwen3-4B Oracle; Think 15.3%

Source: Official EnterpriseOps-Gym leaderboard

#EntryScoreSettingDateSource
1 Claude Opus 4.6 45.9% Oracle Feb 2026 Official leaderboard
2 Claude Sonnet 4.6 42.2% Oracle Feb 2026 Official leaderboard
3 Claude Opus 4.5 39.4% Oracle Nov 2025 Official leaderboard
4 Gemini 3.1 Pro 38.7% Oracle Feb 2026 Official leaderboard
5 Claude Sonnet 4.5 34.1% Oracle Sep 2025 Official leaderboard
6 Gemini 3 Flash 34% Oracle Dec 2025 Official leaderboard
7 GPT-5 30.9% Oracle Aug 2025 Official leaderboard
8 Gemini 3 Pro 29.7% Oracle Nov 2025 Official leaderboard
9 DeepSeek-V3.2 24.2% Oracle Dec 2025 Official leaderboard
10 GPT-5 mini 22.9% Oracle Aug 2025 Official leaderboard
11 Kimi K2 Thinking 22.4% Oracle Nov 2025 Official leaderboard
12 Gemini-2.5-Pro (snapshot unspecified) 20% Oracle Official leaderboard
13 Qwen3-30B 19.1% Oracle; Think Official leaderboard
14 Qwen3-235B 18.1% Oracle; Inst. Official leaderboard
15 Qwen3-4B 15.3% Oracle; Think Apr 2025 Official leaderboard
#ModelScoreSettingDateSource
1 Claude Opus 4.6 45.9% Oracle Feb 2026 Official leaderboard
2 Claude Sonnet 4.6 42.2% Oracle Feb 2026 Official leaderboard
3 Claude Opus 4.5 39.4% Oracle Nov 2025 Official leaderboard
4 Gemini 3.1 Pro 38.7% Oracle Feb 2026 Official leaderboard
5 Claude Sonnet 4.5 34.1% Oracle Sep 2025 Official leaderboard
6 Gemini 3 Flash 34% Oracle Dec 2025 Official leaderboard
7 GPT-5 30.9% Oracle Aug 2025 Official leaderboard
8 Gemini 3 Pro 29.7% Oracle Nov 2025 Official leaderboard
9 DeepSeek-V3.2 24.2% Oracle Dec 2025 Official leaderboard
10 GPT-5 mini 22.9% Oracle Aug 2025 Official leaderboard
11 Kimi K2 Thinking 22.4% Oracle Nov 2025 Official leaderboard
12 Qwen3-4B 15.3% Oracle; Think Apr 2025 Official leaderboard
#ModelScoreSettingDateSource
1 Inkling 2 Small 46.8% Nov 2026 Lab model card
2 Claude Opus 4.6 45.9% Oracle Feb 2026 Official leaderboard
3 Claude Sonnet 4.6 42.2% Oracle Feb 2026 Official leaderboard
4 Inkling 2 Mini 42% Nov 2026 Lab model card
5 Claude Opus 4.5 39.4% Oracle Nov 2025 Official leaderboard
6 Gemini 3.1 Pro 38.7% Oracle Feb 2026 Official leaderboard
7 Claude Sonnet 4.5 34.1% Oracle Sep 2025 Official leaderboard
8 Gemini 3 Flash 34% Oracle Dec 2025 Official leaderboard
9 Spark-X2.6-4B 33.6% Oct 2026 Lab model card
10 GPT-5 30.9% Oracle Aug 2025 Official leaderboard
11 Gemini 3 Pro 29.7% Oracle Nov 2025 Official leaderboard
12 LFM3-1.7B 28.2% Oct 2026 Lab model card
13 DeepSeek-V3.2 24.2% Oracle Dec 2025 Official leaderboard
14 GPT-5 mini 22.9% Oracle Aug 2025 Official leaderboard
15 Kimi K2 Thinking 22.4% Oracle Nov 2025 Official leaderboard
16 Qwen3-4B 15.3% Oracle; Think Apr 2025 Official leaderboard

Labs report scores for 4 models in their cards. Switch the scores above to rank them together.

Rankings by part

We read 8 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as Teams, Email and Calendar. Claude Opus 4.6 leads 5 of them, and 2 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.

Teams

12 models we track have a score here.

  1. 1 Claude Opus 4.6 52%
  2. 2 Claude Sonnet 4.5 ↑3 51%
  3. 3 Claude Opus 4.5 50%
  4. 4 Gemini 3 Flash ↑2 47.3%
  5. 5 Claude Sonnet 4.6 ↓3 47%
  6. 6 Gemini 3.1 Pro ↓2 46%
  7. 7 Gemini 3 Pro ↑1 43%
  8. 8 DeepSeek-V3.2 ↑1 35.7%
  9. 9 Kimi K2 Thinking ↑2 30%
  10. 10 GPT-5 ↓3 26.3%

Number one over time 2 models have held number one with claims included since its release in March 2026

Lab claims count from the day the card came out, independent runs from the model’s release. We draw one row each for the benchmark’s own leaderboard and what labs claim in their cards.

Anthropic iFlytek Thinking Machines
Own leaderboard Not tested since Mar 2026
Lab claims 2 leaders

Sources: Official EnterpriseOps-Gym leaderboard and labs’ model cards

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 Reported by the lab in its model card Best so far
0% 20% 40% 60% 80% 100% 2025 2026 2027 Near the ceiling (85%) Inkling 2 Small Claude Opus 4.6 Claude Opus 4.5 Claude Sonnet 4.5 Qwen3-4B 0% 20% 40% 60% 80% 100% 2025 2026 2027 Near the ceiling (85%) Inkling 2 Small Claude Opus 4.5

Sources: Official EnterpriseOps-Gym leaderboard and labs’ model cards

3 labs put it in a launch post in the last six months

Launch posts from every lab that has listed it, one tick per post. Tall colored ticks include it, and short gray ones leave it out.

iFlytek
Liquid AI
Thinking Machines
DateLaunch postScoreAs the post writes it
Nov 20, 2026 Inkling 2 Mini 42 EnterpriseOps-Gym
Nov 5, 2026 Inkling 2 Small 46.8 EnterpriseOps-Gym
Oct 27, 2026 LFM3-1.7B 28.2 EnterpriseOps-Gym
Oct 22, 2026 Spark-X2.6-4B 33.6 EnterpriseOps-Gym

One system card reports it

Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.

DateDocumentScoreAs the document writes itSection
Nov 20, 2026 Inkling 2 Mini model card 42 EnterpriseOps-Gym Capabilities

Versions

More in Tool use

All 265 in Tool use
  • AutomationBench In use Agent completes end-to-end business workflows inside a simulated company exposing dozens of REST APIs. Graded on resulting state. Gemini 4 Neon 56.3%
  • MCP Atlas In use Multi-step workflows over real MCP servers. Scores tool-use competency on realistic tasks. Gemini 4 Argon 90.7%
  • BFCL V4 In use Models call functions and tools, including multi-turn agentic tasks with web search and memory. Checked against expected calls and outcomes. Qwen3.7-Max 75%
  • τ³-bench Banking In use Agents act as banking customer-service reps over many turns with a simulated user, following policy. Pass^1 on the final database state. Qwen 3.8 Max 55.1%
  • Toolathlon-Verified In use 108 long-horizon tasks across 32 apps through about 600 MCP tools. Execution-based checkers inspect final environment state. GLM-5.3-Flash 78.4%
  • DeepSearchQA In use Questions that need multi-step search and synthesis. Final answers graded against references. Kimi K3 95%

Data audit

All 10 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
  • Pass Independent runs Scores from the official leaderboard (15 entries), launch posts (4 scores) and system cards (1 score).
  • Pass Models scored 16 models we track, plus 3 other entries on its leaderboard.
  • Pass Newest score The newest is from November 2026, for Inkling 2 Mini.
  • Note Frontier models None of the top 10 has a score, which is fine for a retired test. The strongest tested is Claude Opus 4.6 (#39).
  • Pass Headroom The best is 46.8%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on github.com: 12 rows for models we track and 3 other entries at the last read.
  • Pass Last read November 18th, 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