Atlas

Benchmarks

TCSExam solver

GitHub

TCSExam's solver track: 30 solver-engineering problems from competitions and instance libraries, judged on hidden instances. Each instance scores 100Φ((a−μ)/σ) against three fixed Kimi K3 runs, so 50 matches Kimi K3's mean.

Models
8
Top score
64.7
Leader
GPT-6 Astra
Updated
Nov 5, 2026
Resolution
—
Saturation
—
Signal
92ndpercentile
205 benchmarks from 0.8 to 3.6219 benchmarks from 3.6 to 6.4134 benchmarks from 6.4 to 9.393 benchmarks from 9.3 to 12.155 benchmarks from 12.1 to 14.925 benchmarks from 14.9 to 17.713 benchmarks from 17.7 to 20.55 benchmarks from 20.5 to 23.37 benchmarks from 23.3 to 26.27 benchmarks from 26.2 to 29.01 benchmark from 29.0 to 31.83 benchmarks from 31.8 to 34.66 benchmarks from 34.6 to 37.43 benchmarks from 37.4 to 40.21 benchmark from 40.2 to 43.11 benchmark from 45.9 to 48.71 benchmark from 65.6 to 68.4

61st strongest of 779

Signal-to-noise +16.16 for the TCSExam family

Difficulty
Not estimated

Another benchmark represents this family in the fit.

8 dated models · 4 frontier modelsClosed weightsOpen weights
Data: Moonshot AIatlas.village.dev
#ModelSettingScoreSource
1GPT-6 Astramax · Codex
64.7
TCSExam paper Table 2 (Full, GPT-6 Astra)
2Kimi K3.1 Promax · Kimi Code
54.1
TCSExam paper Table 2 (Full, Kimi K3.1 Pro)
3DeepSeek V4.1 Flash 1017max · Claude Code
53.1
TCSExam paper Table 2 (Full, DeepSeek V4.1 Flash 1017)
4Qwen3.8 Max (0902)max · Claude Code
49.2
TCSExam paper Table 2 (Full, Qwen3.8 Max (0902))
5GLM-5.3max · Claude Code
49.1
TCSExam paper Table 2 (Full, GLM-5.3)
6Kimi K3max · Kimi Code
48.7
TCSExam paper Table 2 (Full, Kimi K3)
7DeepSeek V4.1 Flashmax · Claude Code
48.3
TCSExam paper Table 2 (Full, DeepSeek V4.1 Flash 0910)
8Grok 4.6x-high · Grok Build
43.8
TCSExam paper Table 2 (Full, Grok 4.6)
Loading Atlas data…