Benchmark Almanac

MCPMark Verified

Kimi K3 96.1% leads since July 2026
Not in model cards

Agents complete 127 tasks through version-pinned MCP servers. Stabilized verifiers score single-run task success.

The authors deprecate earlier task versions and state that their results are not directly comparable.

Released
June 2026
Kind
Apps and MCP servers
Level
University
Scored
Checked automatically
Answers
Agent task
Tasks
All public
Strongest model tested
Claude Fable 5 #11 on the capability index
Last model added
July 2026
Scores refreshed
November 12th, 2026
8
Models scored on its own leaderboard

Kimi K3 leads with 96.1% among models released by July 2026

Best score per model from the benchmark's own leaderboard, read on November 12th, 2026. Where a model ran with several settings or harnesses, we keep its best.

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

  1. 96.1 Kimi K3
  2. 92.9 GPT-5.5
  3. 92.9 GPT-5.6 Sol
  4. 86.6 Claude Fable 5
  5. 81.9 Kimi K2.7 Code
  6. 76.4 Claude Opus 4.8
  7. 72.8 Kimi K2.6
  8. 71.7 DeepSeek-V4-Pro
  1. 1 Kimi K3 max 96.1%
  2. 2 GPT-5.5 xhigh 92.9%
  3. 3 GPT-5.6 Sol max 92.9%
  4. 4 Claude Fable 5 max 86.6%
  5. 5 Kimi K2.7 Code 81.9%
  6. 6 Claude Opus 4.8 max 76.4%
  7. 7 Kimi K2.6 72.8%
  8. 8 DeepSeek-V4-Pro max 71.7%

Source: Official MCPMark Verified leaderboard

#ModelScoreSettingDateSource
1 Kimi K3 96.1% max Jul 2026 Official leaderboard
2 GPT-5.5 92.9% xhigh Apr 2026 Official leaderboard
3 GPT-5.6 Sol 92.9% max Jul 2026 Official leaderboard
4 Claude Fable 5 86.6% max Jun 2026 Official leaderboard
5 Kimi K2.7 Code 81.9% Jun 2026 Official leaderboard
6 Claude Opus 4.8 76.4% max May 2026 Official leaderboard
7 Kimi K2.6 72.8% Apr 2026 Official leaderboard
8 DeepSeek-V4-Pro 71.7% max Apr 2026 Official leaderboard

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

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

OpenAI Moonshot AI
Own leaderboard Passed 85% in Jun 2026, by GPT-5.5
Kimi K3 96.1%

Source: Official MCPMark Verified leaderboard

GPT-5.5 on the official leaderboard passed 85% in April 2026

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

On the benchmark’s own leaderboard Best so far
0% 20% 40% 60% 80% 100% 2026 Near the ceiling (85%) Kimi K3 GPT-5.5 Kimi K2.6 0% 20% 40% 60% 80% 100% 2026 Near the ceiling (85%) Kimi K3 Kimi K2.6

Source: Official MCPMark Verified leaderboard

No lab model card reports it yet

We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark.

Versions

We switch the top row to each new version when it comes out. Open it for one row per version. An old version stays in play while people still test it, so whoever first takes it past 85% gets credit for closing it out.

DeepSeek ByteDance Seed Moonshot AI Qwen OpenAI
MCPMark Verified Passed 85% in Jun 2026, by GPT-5.5
Kimi K3 96.1%
MCPMark Passed 85% in Jul 2026, by Kimi K3
Kimi K3 94.5%
MCPMark Verified Passed 85% in Jun 2026, by GPT-5.5
Kimi K3 96.1%

Sources: Official leaderboards and labs’ model cards

Replaces MCPMark

More in Tool use

All 144 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 Carbon 53.9%
  • 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%