Benchmark Almanac

OmniSch

Qwen3.5-122B-A10B 0.28 leads since March 2026
Not in model cards

Models convert circuit schematics into graphs of components, pins, and connections. A second task lets models search schematic images for individual features.

The authors compare nine general-purpose models in each of two tables on private data. The released dataset contains 1,854 rendered EAGLE sheets.

Released
March 2026
Kind
Electronics
Level
University
Scored
Checked automatically
Answers
Open-ended
Tasks
Public, plus a private test set
Measures
Capability
Language
English
In and out
Text, Images → Text
Interaction
Single turn
Refresh
One fixed set
Run by
A host
In views
Vision
Strongest model tested
Gemini 3.1 Pro #40 on the capability index
Last model added
February 2026
Scores refreshed
November 20th, 2026
9
Entries on its own leaderboard, 8 models we track

Qwen3.5-122B-A10B leads with 0.28 among models released by February 2026 Qwen3.5-122B-A10B leads with 0.28 among models released by February 2026

Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart. 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.

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 Qwen3.5-122B-A10B FewShot. 0.28
  2. 2 Claude Sonnet 4.6 FewShot. 0.26
  3. 3 Gemini 2.5 Flash Sep 2025, w/o Txt. 0.21
  4. 4 Gemini 3.1 Pro Real 0.18
  5. 5 GPT-5.2 Synth. 0.17
  6. 6 Ministral 3 14B FewShot. 0.12
  7. 7 GPT-5 mini FewShot. 0.11
  8. 8 Llama 4 Maverick FewShot. 0.08
  9. 9 Qwen3-VL 8B FewShot. 0.00

Source: Official OmniSch leaderboard

#EntryScoreSettingDateSource
1 Qwen3.5-122B-A10B 0.28 FewShot. Feb 2026 Official leaderboard
2 Claude Sonnet 4.6 0.26 FewShot. Feb 2026 Official leaderboard
3 Gemini 2.5 Flash Sep 2025 0.21 w/o Txt. Sep 2025 Official leaderboard
4 Gemini 3.1 Pro 0.18 Real Feb 2026 Official leaderboard
5 GPT-5.2 0.17 Synth. Dec 2025 Official leaderboard
6 Ministral 3 14B 0.12 FewShot. Dec 2025 Official leaderboard
7 GPT-5 mini 0.11 FewShot. Aug 2025 Official leaderboard
8 Llama 4 Maverick 0.08 FewShot. Apr 2025 Official leaderboard
9 Qwen3-VL 8B open weights, 8B 0.00 FewShot. Official leaderboard
#ModelScoreSettingDateSource
1 Qwen3.5-122B-A10B 0.28 FewShot. Feb 2026 Official leaderboard
2 Claude Sonnet 4.6 0.26 FewShot. Feb 2026 Official leaderboard
3 Gemini 2.5 Flash Sep 2025 0.21 w/o Txt. Sep 2025 Official leaderboard
4 Gemini 3.1 Pro 0.18 Real Feb 2026 Official leaderboard
5 GPT-5.2 0.17 Synth. Dec 2025 Official leaderboard
6 Ministral 3 14B 0.12 FewShot. Dec 2025 Official leaderboard
7 GPT-5 mini 0.11 FewShot. Aug 2025 Official leaderboard
8 Llama 4 Maverick 0.08 FewShot. Apr 2025 Official leaderboard

Scores at each effort level

Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.

Lowest effort to highest, light to dark
Qwen3.5-122B-A10B
Real w/o Txt. Synth. FewShot. Real Low R.
0.28
Claude Sonnet 4.6
Real w/o Txt. FewShot. Real Low R.
0.26
Gemini 2.5 Flash
Synth. Real Real Low R.
0.21
Gemini 3.1 Pro
w/o Txt. Synth. FewShot. Real Low R.
0.18
GPT-5.2
w/o Txt. FewShot. Synth. Real Low R.
0.17
Ministral 3 14B
w/o Txt. Real FewShot. Real Low R.
0.12
GPT-5 mini
w/o Txt. Synth. FewShot. Real Low R.
0.11
Llama 4 Maverick
w/o Txt. Real FewShot. Real Low R.
0.08

Source: Official OmniSch leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Qwen3.5-122B-A10B FewShot. 0.28
Claude Sonnet 4.6 FewShot. 0.26
Qwen3.5-122B-A10B Synth. 0.25
Gemini 2.5 Flash Sep 2025 w/o Txt. 0.21
Gemini 2.5 Flash Sep 2025 FewShot. 0.21
Gemini 2.5 Flash Sep 2025 Real 0.19
Qwen3.5-122B-A10B w/o Txt. 0.18
Gemini 3.1 Pro Real 0.18
Claude Sonnet 4.6 w/o Txt. 0.17
GPT-5.2 Synth. 0.17
Gemini 3.1 Pro FewShot. 0.17
GPT-5.2 Real 0.15
GPT-5.2 FewShot. 0.14
Claude Sonnet 4.6 Synth. 0.14
Gemini 3.1 Pro Synth. 0.13
Gemini 2.5 Flash Sep 2025 Real Low R. 0.13
Gemini 2.5 Flash Sep 2025 Synth. 0.13
Claude Sonnet 4.6 Real 0.12
Ministral 3 14B FewShot. 0.12
Qwen3.5-122B-A10B Real 0.12
GPT-5 mini FewShot. 0.11
Ministral 3 14B Synth. 0.09
Qwen3.5-122B-A10B Real Low R. 0.09
Gemini 3.1 Pro Real Low R. 0.08
Claude Sonnet 4.6 Real Low R. 0.08
Ministral 3 14B Real 0.08
Llama 4 Maverick FewShot. 0.08
GPT-5 mini Real 0.07
GPT-5 mini Synth. 0.07
Llama 4 Maverick Synth. 0.06
GPT-5.2 Real Low R. 0.05
Llama 4 Maverick Real 0.05
Gemini 3.1 Pro w/o Txt. 0.04
Ministral 3 14B Real Low R. 0.03
GPT-5 mini Real Low R. 0.02
Llama 4 Maverick Real Low R. 0.02
GPT-5.2 w/o Txt. 0.01
Ministral 3 14B w/o Txt. 0.01
GPT-5 mini w/o Txt. 0.00
Llama 4 Maverick w/o Txt. 0.00

Scores at each effort level

Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.

Lowest effort to highest, light to dark
Qwen3.5-122B-A10B
Real w/o Txt. Synth. FewShot. Real Low R.
0.28
Claude Sonnet 4.6
Real w/o Txt. FewShot. Real Low R.
0.26
Gemini 2.5 Flash
Synth. Real Real Low R.
0.21
Gemini 3.1 Pro
w/o Txt. Synth. FewShot. Real Low R.
0.18
GPT-5.2
w/o Txt. FewShot. Synth. Real Low R.
0.17
Ministral 3 14B
w/o Txt. Real FewShot. Real Low R.
0.12
GPT-5 mini
w/o Txt. Synth. FewShot. Real Low R.
0.11
Llama 4 Maverick
w/o Txt. Real FewShot. Real Low R.
0.08

Source: Official OmniSch leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Qwen3.5-122B-A10B FewShot. 0.28
Claude Sonnet 4.6 FewShot. 0.26
Qwen3.5-122B-A10B Synth. 0.25
Gemini 2.5 Flash Sep 2025 w/o Txt. 0.21
Gemini 2.5 Flash Sep 2025 FewShot. 0.21
Gemini 2.5 Flash Sep 2025 Real 0.19
Qwen3.5-122B-A10B w/o Txt. 0.18
Gemini 3.1 Pro Real 0.18
Claude Sonnet 4.6 w/o Txt. 0.17
GPT-5.2 Synth. 0.17
Gemini 3.1 Pro FewShot. 0.17
GPT-5.2 Real 0.15
GPT-5.2 FewShot. 0.14
Claude Sonnet 4.6 Synth. 0.14
Gemini 3.1 Pro Synth. 0.13
Gemini 2.5 Flash Sep 2025 Real Low R. 0.13
Gemini 2.5 Flash Sep 2025 Synth. 0.13
Claude Sonnet 4.6 Real 0.12
Ministral 3 14B FewShot. 0.12
Qwen3.5-122B-A10B Real 0.12
GPT-5 mini FewShot. 0.11
Ministral 3 14B Synth. 0.09
Qwen3.5-122B-A10B Real Low R. 0.09
Gemini 3.1 Pro Real Low R. 0.08
Claude Sonnet 4.6 Real Low R. 0.08
Ministral 3 14B Real 0.08
Llama 4 Maverick FewShot. 0.08
GPT-5 mini Real 0.07
GPT-5 mini Synth. 0.07
Llama 4 Maverick Synth. 0.06
GPT-5.2 Real Low R. 0.05
Llama 4 Maverick Real 0.05
Gemini 3.1 Pro w/o Txt. 0.04
Ministral 3 14B Real Low R. 0.03
GPT-5 mini Real Low R. 0.02
Llama 4 Maverick Real Low R. 0.02
GPT-5.2 w/o Txt. 0.01
Ministral 3 14B w/o Txt. 0.01
GPT-5 mini w/o Txt. 0.00
Llama 4 Maverick w/o Txt. 0.00

Number one over time

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

Qwen
Own leaderboard Not tested since Mar 2026

Source: Official OmniSch leaderboard

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
0.00 0.10 0.20 0.30 2025 2026 0.00 0.10 0.20 0.30 2025 2026

Source: Official OmniSch leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

More in Engineering and CAD

All 112 in Engineering and CAD
  • EEBench In use 13 electrical-engineering circuit and PCB design tasks graded by deterministic build checks, simulation and BOM constraints. No LLM judge. Gemini 4 Carbon 83.9%
  • 3DCodeBench In use Agentic procedural 3D modeling via code: author engine-ready 3D assets through software APIs. Graded for executability and shape fidelity. LongCat-2.5 55.4%
  • CADGenBench Only in system cards Agents construct CAD models from design prompts. The generation split is graded for executability and geometric correctness. Claude Fable 5 45.1%
  • PartBench Only in system cards Parametric mechanical part design and CAD-style generation, graded for geometric and manufacturing fidelity. Grok 4.6 58.5%
  • Razavi-Bench Not in model cards Models explain what happens in analog circuits and why. A language model awards up to four points per answer against a reference and a rubric. Gemini 4 Carbon 98.3%
  • CircuitRubric Not in model cards Models write analog circuit netlists from design requests. A program compares transistor connections and relative device sizes with accepted reference circuits. Claude Opus 4.8 96%

Data audit

8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 9 months ago, in February 2026, for Qwen3.5-122B-A10B. Every benchmark

Catalog record 2 pass
  • Pass Task access Public tasks, plus a private test set.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 2 pass 2 notes
  • Pass Independent runs Scores from the official leaderboard (9 entries).
  • Pass Models scored 8 models we track, plus 1 other entry on its leaderboard.
  • Warning Newest score The newest came 9 months ago, in February 2026, for Qwen3.5-122B-A10B.
  • Note Frontier models None of the top 10 has a score. The strongest tested is Gemini 3.1 Pro (#40).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on omnisch.com: 40 rows for models we track and 1 other entry at the last read.
  • Pass Last read November 20th, 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