Note: Overall leaderboard rankings may not reflect true model quality — individual benchmarks give a clearer picture. ARC-Challenge MMLU GPQA GSM8K Artificial Analysis Intelligence Index v4.1
← Back to leaderboard

reflection: beam-1.0

4 benchmarks
Companyreflection

Leaderboard position

#114
Overall rank (of 1121 models)
2205
Elo rating
4
Benchmarks
Top 10.2%
Percentile
#1
In reflection (of 1 models)
Above: NON AI - НЕ ИИ: Human baseline - Человеческий результат #113 Below: reve: reve-2.0 #115

Benchmark highlights

Best benchmark
SWE-bench Pro 66.1
Worst benchmark
Artificial Analysis Intelligence Index v5.0.1 18

Nearest competitor from another company

The model from another company with the most similar benchmark scores.

appeared as nearest in 2 of 4 benchmarks

ifm: k2-horizon-375b-a23b
2/4 50.0%
Model Votes % of benchmarks
ifm: k2-horizon-375b-a23b 2 50.0%
Qwen: Qwen3.7 Max 1 25.0%
Tencent: hy4-preview 1 25.0%

Benchmarks

SWE-bench Pro 66.1 Artificial Analysis Intelligence Index v4.3.2 30 Artifical Analysis Intelligence Index v5.0 18 Artificial Analysis Intelligence Index v5.0.1 18

Predicted scores

Estimated from 25 similar models and the global Elo rating. These are projections, not measured results.

Benchmark Predicted Confidence Based on
AIME 2026 ~99.2 60% Meta: muse-glimmer-30b
GPQA Diamond ~87.8 60% Qwen: qwen3.8-27b, Meta: muse-spark-open-816b-a42b, Meta: muse-glimmer-30b
IFBench ~75.69 60% Qwen: qwen3.8-27b, Meta: muse-glimmer-30b
Terminal-Bench v2.1 ~68.32 60% Qwen: qwen3.8-27b, Meta: muse-spark-open-816b-a42b, Z.ai: glm-5.3
SWE-bench Verified ~81.56 60% Meta: muse-glimmer-30b
SciCode ~48.02 60% Meta: muse-spark-open-816b-a42b, Meta: muse-glimmer-30b, OpenAI: gpt-6.1-sol
AA-LCR ~69.48 60% Mistral AI: Mistral Large 4 (Le Chonk), Meta: muse-glimmer-30b, OpenAI: gpt-6-astra
LiveCodeBench v6 ~86.78 60% Qwen: qwen3.8-27b
MMMU-Pro ~78.31 60% Mistral AI: Mistral Large 4 (Le Chonk), Meta: muse-glimmer-30b, OpenAI: gpt-6.1-sol
BrowseComp ~74.61 60% OpenAI: gpt-6-astra
Artificial Analysis Intelligence Index v4.1.1 (Maximum Reasoning) ~44.35 60% Thinking Machines Lab: inkling, Meta: muse-glimmer-30b, Google: gemini-3.8-flash
ScreenSpot-Pro ~79 60% Meta: muse-glimmer-30b
Artificial Analysis Intelligence Index v4.1 (Maximum Reasoning) ~38.25 60% Thinking Machines Lab: inkling
Humanity's Last Exam (no tools) ~35.95 59% Qwen: qwen3.8-27b, DeepSeek: deepseek-v4.1-flash-1017, Anthropic: claude-haiku-5.5
Humanity's Last Exam (with tools) ~48.66 58% Meta: muse-spark-open-816b-a42b, Z.ai: glm-5.3, Anthropic: claude-haiku-5.5
Claw-Eval ~67.55 58% Qwen: qwen3.8-27b
ParseBench ~68.85 58% Qwen: qwen3.8-27b, Anthropic: claude-opus-5.5
AA-Omniscience Accuracy ~30.33 54% Meta: muse-spark-open-816b-a42b, Mistral AI: Mistral Large 4 (Le Chonk), Anthropic: claude-sonnet-5.5
MCP-Atlas ~72.84 52% Meta: muse-spark-open-816b-a42b, Meta: muse-glimmer-30b
Artificial Analysis Intelligence Index v4.3 ~32.6 52% Qwen: qwen3.8-27b, Thinking Machines Lab: inkling, Z.ai: glm-5.3
MATH-500 ~99.2 50% —
AIME 2025 ~100 50% —
HMMT Feb 2026 ~91.22 50% —
MMLU-Pro ~90.53 50% —
IFEval ~96.75 50% —