Clinical Benchmarks

Moonshot AI

Moonshot AI models with results here, each shown on its own board.

Models
4
Results
20
Boards
11 of 14

www.moonshot.ai

Models

ModelKindReleasedBoards
Kimi K3.1 ProModel23 Oct 20261
Kimi K3Model16 Jul 20268
Kimi K2.6Model20 Apr 20265
Kimi K2.5Model27 Jan 20263

Results

One line per result. Tall tick: the board leader.

Clinical reasoning and knowledge

  1. HealthBench Professional
    Artificial Analysis independent run, max reasoning; GPT-5.4 (low) judge, 4 repeats per task; length-adjusted (56.9 raw, 3,250 mean response chars)
    0.532
    Rank 28 of 43Leader GPT-6 Astra (Anthropic run) 0.703
    Independent run
    Measured Oct 2026
  2. MedXpertQA (MM)
    Qwen-run comparison in the Qwen3.5-397B-A17B model card (K2.5-1T-A32B column)
    65.3
    Rank 14 of 22 here, 5 models on the boardLeader GPT-5.6 Sol 81.5
    Independent run
  3. 60.1
    Rank 2 of 8 here, 11 models on the boardLeader GPT-5.6 Sol 60.2
    Official leaderboard
    Measured Aug 2026
  4. 89.7
    Rank 6 of 15 here, 16 models on the boardLeader Claude Opus 5 94.6
    Official leaderboard
    Measured Oct 2026
  5. APEX-1 GP
    Score (avg@4), reasoning max
    64.35
    Rank 4 of 12Leader Claude Opus 5.5 73.04
    Official leaderboard
    Measured Oct 2026

Documentation and coding

  1. 87.96
    Rank 13 of 106 here, 105 models on the boardLeader Claude Opus 5.5 91.43
    Official leaderboard
    Measured Sep 2026
  2. MedScribe (Vals AI)
    model ID kimi/kimi-k2.6; top_p=0.95; max_output_tokens=30000
    78.15
    Rank 63 of 106 here, 105 models on the boardLeader Claude Opus 5.5 91.43
    Official leaderboard
    Measured Sep 2026
  3. MedScribe (Vals AI)
    model ID kimi/kimi-k2.5-thinking; top_p=0.95; max_output_tokens=30000
    76.44
    Rank 73 of 106 here, 105 models on the boardLeader Claude Opus 5.5 91.43
    Official leaderboard
    Measured Sep 2026
  4. MedCode (Vals AI)
    model ID kimi/kimi-k3; temperature=1; max_output_tokens=30000
    48.88
    Rank 25 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  5. MedCode (Vals AI)
    model ID kimi/kimi-k2.6; top_p=0.95; max_output_tokens=30000
    40.14
    Rank 65 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  6. MedCode (Vals AI)
    model ID kimi/kimi-k2.5-thinking; top_p=0.95; max_output_tokens=30000
    39.32
    Rank 66 of 104 here, 103 models on the boardLeader Claude Opus 5 63.57
    Official leaderboard
    Measured Sep 2026
  7. MLCR-AA
    overall pass rate
    38.3
    Rank 3 of 27Leader Claude Sonnet 5.5 (Max, Default Fallback) 41.7
    Independent run
    Measured Nov 2026

EHR and workflow agents

  1. 17.0
    Rank 22 of 28 here, 12 models on the boardLeader Claude Opus 5.5 (max) 68.4
    Official leaderboard
    Measured May 2026
  2. 25.3
    Rank 7 of 43 here, 45 models on the boardLeader erius + claude-opus-5 54.7
    Official leaderboard
    Measured Jul 2026
  3. CHI-Bench
    All Domains pass@1; hermes harness
    15.6
    Rank 21 of 43 here, 45 models on the boardLeader erius + claude-opus-5 54.7
    Official leaderboard
    Measured May 2026
  4. CHI-Bench
    All Domains pass@1; openai-agents harness
    15.1
    Rank 23 of 43 here, 45 models on the boardLeader erius + claude-opus-5 54.7
    Official leaderboard
    Measured May 2026
  5. CHI-Bench
    All Domains pass@1; openclaw harness
    10.2
    Rank 32 of 43 here, 45 models on the boardLeader erius + claude-opus-5 54.7
    Official leaderboard
    Measured May 2026
  6. CHI-Bench
    All Domains pass@1; deepagents harness
    3.1
    Rank 40 of 43 here, 45 models on the boardLeader erius + claude-opus-5 54.7
    Official leaderboard
    Measured May 2026

Safety

  1. 74.0
    Rank 7 of 17 here, 19 models on the boardLeader LiSA 2.5 86.2
    Official leaderboard
    Measured Aug 2026
  2. 59.1
    Rank 15 of 17 here, 19 models on the boardLeader LiSA 2.5 86.2
    Official leaderboard
    Measured Aug 2026