AgentBench / results
← All models

Phi-4 Mini Instruct 3.8B Q4_K_M

1 tested weight format · Explore results, then inspect the answers.

Choose a tested configuration

Get this tested model

Q4_K_M · 2.49 GB download

Links follow the selected quant and configuration.

Setup, memory & provenance · Q4_K_M

Q4_K_M · 2.49 GB of weights · vulkan · context 16384 · GPU layers 99

Small · under 5B · dense · 3.8B total

Memory fit has not been verified for this configuration.

partial

Recorded settings and thinking verification

Temperature 0 · seed 42 · output limit 4096 · reasoning effort unrecorded

RAM labels describe the test computer, not this model’s memory requirement. Speed comparisons keep the 32 GB and 128 GB runs separate.

Runtime build: unrecorded · Local · test machine with 32 GB RAM · concurrency: 1

Thinking status: ignored switch · probe 2 · scope Saved profile-level probe; exact historical template/build scope is not fully recorded..

Requested off does not prove an absence of reasoning. Saved checks do not establish unrecorded historical template or build details.

Overall across tasks

14.4 / 100

Nine categories, equally weighted · 9/9 eligible

Complete coverage for this setup. Quality only; speed is separate. How the index works →

What does it do well?

Select a category to explore the answers. Bars show what happened on every planned task. Incorrect answers and failures to finish both earn zero; technical problems and untested tasks stay unscored.

Exact scores and task times by suite

Each suite has its own metric and fixed task bank. The balanced index weights categories equally; individual scores and evidence remain available below.

Test suiteThinking offDetails
Everyday coding tasksTask pass rate
Pass rate
0/16 · 0%
Answers completed
16/16 · 100%
Correct / completed
0/16 · 0%
16 incorrect
16/16 scored3.9s median task time · 16 measured
Larger practical tasksMean task credit
Average credit
0.0%
Full-credit rate
0/11 · 0%
Answers completed
11/11 · 100%
Full credit / completed
0/11 · 0%
11 incorrect
11/11 scored4.8s median task time · 11 measured
Logic and number puzzlesMean puzzle credit
Average credit
26.0%
Full-credit rate
6/34 · 18%
Answers completed
34/34 · 100%
Full credit / completed
6/34 · 18%
6 full credit10 partial credit18 incorrect
34/34 scored4.6s median task time · 34 measured
Published hard problemsMean answer credit
Average credit
5.2%
Full-credit rate
0/11 · 0%
Answers completed
11/11 · 100%
Full credit / completed
0/11 · 0%
3 partial credit8 incorrect
11/11 scored6.5s median task time · 11 measured
Following exact instructions · original campaignStrict prompt compliance
Pass rate
14/23 · 61%
Answers completed
23/23 · 100%
Correct / completed
14/23 · 61%
14 full credit9 incorrect
23/23 scored3.8s median task time · 23 measured
Reading without guessingExact-match accuracy
Pass rate
6/16 · 38%
Answers completed
16/16 · 100%
Correct / completed
6/16 · 38%
6 full credit10 incorrect
16/16 scored0.7s median task time · 16 measured
Calculating from a passageNumeric exact-match accuracy
Pass rate
0/7 · 0%
Answers completed
7/7 · 100%
Correct / completed
0/7 · 0%
7 incorrect
7/7 scored1.0s median task time · 7 measured
Reading research papersAnswer token F1
Average credit
0.0%
Full-credit rate
0/8 · 0%
Answers completed
8/8 · 100%
Full credit / completed
0/8 · 0%
8 incorrect
8/8 scored3.9s median task time · 8 measured
Resisting incorrect correctionsAll-branches pass rate
Pass rate
0/12 · 0%
Answers completed
12/12 · 100%
Correct / completed
0/12 · 0%
12 incorrect
12/12 scored0.8s median task time · 12 measured
Following exact instructions · 128 GB campaignStrict prompt complianceNot planned

Model analysis

This analysis covers this configuration across its recorded thinking conditions and test suites.

Consider this Instruct model only for low-stakes wording and formatting tasks you can check yourself. With thinking requested off, it followed most strict instructions, but neither small-project edits nor larger practical tasks succeeded. Research-paper answering and source-challenge tests were also unsuccessful. Short waits do not make a compelling speed/quality choice here: some other tested small dense models followed instructions more accurately with similar or shorter task times on this test machine.

Evidence for this judgment
Following exact instructions · Thinking off
Tasks passed14 of 23 (60.9%)
Test coverage23 of 23 tasks scored
Tasks timed23 tasks
Typical time per task (median)3.8 seconds
Explore these test results →

View this model’s answers, code and tool activity →

Everyday coding tasks · Thinking off
Tasks passed0 of 16 (0.0%)
Test coverage16 of 16 tasks scored
Explore these test results →

View this model’s answers, code and tool activity →

Larger practical tasks · Thinking off
Average score0.0/100 · Mean task credit
Test coverage11 of 11 tasks scored
Task outcomes0 full credit · 0 partial · 11 no credit
Explore these test results →

View this model’s answers, code and tool activity →

Reading research papers · Thinking off
Average score0.0/100 · Answer token F1
Test coverage8 of 8 tasks scored
Task outcomes0 full credit · 0 partial · 8 no credit
Explore these test results →

View this model’s answers, code and tool activity →

Resisting incorrect corrections · Thinking off
Tasks passed0 of 12 (0.0%)
Test coverage12 of 12 tasks scored
Explore these test results →

View this model’s answers, code and tool activity →

Following exact instructions
Thinking off. Small · under 5B · dense. Rank 6 of 9 in this recorded comparison.

Speed: Following exact instructions
Thinking off. Matching test-machine label and execution settings; spans model sizes. Timings include unsuccessful tasks, so read them alongside quality.

Speed and task success: Following exact instructionsThinking off · Matching test-machine label and settings · Across model sizes

Shorter task time is better only when the result is useful. These times include failures, tools and grading. Token generation speed measures a different part of the work.

Compare timings and outcomes
ModelTypical task timeGeneration speedTasks passed
LFM2.5 1.2B Q4_01.7 seconds23 of 23 tasks timed319.0 tokens/s20 of 2323 of 23 scored
Qwen3.5 0.8B Q4_02.9 seconds23 of 23 tasks timed248.5 tokens/s8 of 2323 of 23 scored
Ling 3.0 Tiny Q4_03.3 seconds23 of 23 tasks timed138.9 tokens/s14 of 2323 of 23 scored
Mellum 2.1 12B-A2.5B Thinking3.6 seconds23 of 23 tasks timed140.2 tokens/s14 of 2323 of 23 scored
Granite 4.0 Micro 3.4B Q4_03.6 seconds23 of 23 tasks timed111.7 tokens/s16 of 2323 of 23 scored
SmolLM3 3B Q4_K_M3.7 seconds23 of 23 tasks timed129.6 tokens/s12 of 2323 of 23 scored
Phi-4 Mini Instruct 3.8B Q4_K_MThis model3.8 seconds23 of 23 tasks timed108.7 tokens/s14 of 2323 of 23 scored
Nemotron Nano 9B v2 Q4_04.9 seconds23 of 23 tasks timed53.6 tokens/s19 of 2323 of 23 scored
Qwen2.5 Coder 7B Instruct Q4_K_M5.0 seconds23 of 23 tasks timed65.1 tokens/s13 of 2323 of 23 scored
Granite 4.2 8B Q4_06.9 seconds23 of 23 tasks timed57.7 tokens/s18 of 2323 of 23 scored
Ornith 1.5 9B Q4_K_M7.3 seconds23 of 23 tasks timed54.8 tokens/s12 of 2323 of 23 scored
Gemma 4 12B QAT Q4_09.4 seconds23 of 23 tasks timed37.8 tokens/s20 of 2323 of 23 scored
Ministral 3 8B Instruct Q4_K_M10.2 seconds23 of 23 tasks timed54.2 tokens/s14 of 2323 of 23 scored
Phi-4 Mini Reasoning 3.8B Q4_K_M15.2 seconds23 of 23 tasks timed105.8 tokens/s8 of 2323 of 23 scored

Results from this test setup, not a promise for your computer. Hardware, software, tokenizers and answer length can change the ordering. Historical runtime builds were not always captured. Comparison saved with the 2026-10-22 assessment.

Test machine: AMD Ryzen 5 5600X (6 cores / 12 threads) · AMD Radeon RX 6700 XT (12 GB VRAM) · 32 GB or 128 GB system RAM, never mixed in one timing group. Full test machine details

Summary generated by GPT Astra · · Edited by Codex; agent reviewed by Codex · publication authorized by user

About this summary
Analyst provider
openai
Generation route
Codex CLI · codex-cli 0.163.0 · ChatGPT account
Requested model
gpt-6-astra
Returned model
Not supplied; identity is recorded from the request only.
Exact provider snapshot
Exact provider snapshot not supplied.
Reasoning effort
high
Prompt version
model-summary-reader-v8
Generation time
2026-10-22T05:57:42.002504Z
Review time
2026-10-22T06:50:41.510511+00:00
Evidence reviewer
Codex (agent)
Publication authorization
User · 2026-10-22T05:55:43.625712+00:00 · Go ahead and complete the remaining work. Spin up agents: luna, terra, sol, astra, as needed to get it done. Narrative summaries: these are the astra generated summaries? Go ahead and get those completed as well. Once all of this is done, you may push the completed result.
Originating evidence edition
dataset-5a6f96ec56c3d240cc42fa23147c1495da1be71f3ef0722c0e5f8c4f5e126a83
Analysis input hash
c9ec1b7e6f9e068d03d9c0b315ddecf70ade4093da45339c02efa2a4f8cac76d
Summary revision
summary-b89faced84eea507f6dc8a6c1b400e10332427f66725e191caf9065a00929eed

Generation and editorial authorship remain attached to this revision. Changing the default generator does not change this attribution.

Good at

Following wording and formatting rules is a usable but limited capability: most requests passed, though a substantial minority did not. Consider it for low-stakes text tasks only when you can check important constraints afterward.

Evidence for this judgment
Following exact instructions · Thinking off
Tasks passed14 of 23 (60.9%)
Test coverage23 of 23 tasks scored
Incorrect completed answers9 tasks
Explore these test results →

View this model’s answers, code and tool activity →

Struggles with

Project work is not a supported use in this setup: neither everyday coding nor larger practical tasks produced a successful result. The weakness extends to small edits, not just longer tasks with tools.

Evidence for this judgment
Everyday coding tasks · Thinking off
Tasks passed0 of 16 (0.0%)
Test coverage16 of 16 tasks scored
Incorrect completed answers16 tasks
Explore these test results →

View this model’s answers, code and tool activity →

Larger practical tasks · Thinking off
Average score0.0/100 · Mean task credit
Test coverage11 of 11 tasks scored
Task outcomes0 full credit · 0 partial · 11 no credit
Incorrect completed answers11 tasks
Explore these test results →

View this model’s answers, code and tool activity →

Source-backed answering was very weak: passage calculations and research-paper answers were unsuccessful, and no source-challenge test passed all checks for retaining and updating answers. Ordinary passage questions also had more misses than exact matches.

Evidence for this judgment
Calculating from a passage · Thinking off
Tasks passed0 of 7 (0.0%)
Test coverage7 of 7 tasks scored
Explore these test results →

View this model’s answers, code and tool activity →

Reading research papers · Thinking off
Average score0.0/100 · Answer token F1
Test coverage8 of 8 tasks scored
Task outcomes0 full credit · 0 partial · 8 no credit
Explore these test results →

View this model’s answers, code and tool activity →

Resisting incorrect corrections · Thinking off
Tasks passed0 of 12 (0.0%)
Test coverage12 of 12 tasks scored
Explore these test results →

View this model’s answers, code and tool activity →

Reading without guessing · Thinking off
Tasks passed6 of 16 (37.5%)
Test coverage16 of 16 tasks scored
Full-credit answers6 tasks
Incorrect completed answers10 tasks
Explore these test results →

View this model’s answers, code and tool activity →

Compared with other models

With thinking requested off, instruction-following offered no distinctive speed advantage among small dense models. LFM2.5 1.2B Q4_0 led that class on accuracy and finished substantially sooner; Granite 4.0 Micro 3.4B Q4_0 was also more accurate with a similar task time in these runs.

Evidence for this judgment
Following exact instructions · Thinking off
Tasks passed14 of 23 (60.9%)
Test coverage23 of 23 tasks scored
Tasks timed23 tasks
Typical time per task (median)3.8 seconds
Explore these test results →

View this model’s answers, code and tool activity →

Following exact instructions
Thinking off. Small · under 5B · dense. Rank 6 of 9 in this recorded comparison.

Speed: Following exact instructions
Thinking off. Matching test-machine label and execution settings; spans model sizes. Timings include unsuccessful tasks, so read them alongside quality.

Following exact instructionsThinking off · Small · under 5B · dense

Keep to rules about wording, length, language and formatting.

Strict prompt compliance · 0–100% · higher is better
23 of 23 tasks scored for each model.

Gaps are relative to this model. Treat small gaps cautiously.

  1. #1 LFM2.5 1.2B Q4_01.2B total parameters87.0%
    6 tasks more passed
  2. #2 Qwen3.5 4B4.2B total parameters82.6%
    5 tasks more passed
  3. =3 Gemma 4 E2B Q4_04.6B total parameters69.6%
    2 tasks more passed
  4. =3 Granite 4.0 Micro 3.4B Q4_03.4B total parameters69.6%
    2 tasks more passed
  5. #5 Qwen3.5 2B Q4_K_M1.9B total parameters65.2%
    1 task more passed
  6. #6 Phi-4 Mini Instruct 3.8B Q4_K_M3.8B total parameters60.9%
    This model
  7. #7 SmolLM3 3B Q4_K_M3.1B total parameters52.2%
    2 tasks fewer passed
  8. =8 Phi-4 Mini Reasoning 3.8B Q4_K_M3.8B total parameters34.8%
    6 tasks fewer passed
  9. =8 Qwen3.5 0.8B Q4_00.8B total parameters34.8%
    6 tasks fewer passed
How this comparison works

Share of tasks that passed the recorded checks. Family comparisons show the labeled Instruct and reasoning versions. Other peer comparisons use the same requested thinking setting, not necessarily each model’s best mode. Small differences on these short tests may not hold on other tasks. Peer membership varies by test coverage and compatible settings.

Comparison saved with the 2026-10-22 assessment. Task-specific results.

With thinking requested off, small-project coding trailed every other tested small dense model, with a large gap to the leaders. Puzzle and difficult-reasoning scores also fell toward the lower end of that class.

Evidence for this judgment
Everyday coding tasks · Thinking off
Tasks passed0 of 16 (0.0%)
Test coverage16 of 16 tasks scored
Explore these test results →

View this model’s answers, code and tool activity →

Logic and number puzzles · Thinking off
Average score26.0/100 · Mean puzzle credit
Test coverage34 of 34 tasks scored
Task outcomes6 full credit · 10 partial · 18 no credit
Explore these test results →

View this model’s answers, code and tool activity →

Published hard problems · Thinking off
Average score5.2/100 · Mean answer credit
Test coverage11 of 11 tasks scored
Task outcomes0 full credit · 3 partial · 8 no credit
Explore these test results →

View this model’s answers, code and tool activity →

Everyday coding tasks
Thinking off. Small · under 5B · dense. Rank 8 of 8 in this recorded comparison.

Logic and number puzzles
Thinking off. Small · under 5B · dense. Rank 7 of 9 in this recorded comparison.

Published hard problems
Thinking off. Small · under 5B · dense. Rank 7 of 9 in this recorded comparison.

Everyday coding tasksThinking off · Small · under 5B · dense

Find information in a small project, edit files and fix bugs.

Task pass rate · 0–100% · higher is better
16 of 16 tasks scored for each model.

Gaps are relative to this model. Treat small gaps cautiously.

  1. #1 Qwen3.5 2B Q4_K_M1.9B total parameters100.0%
    16 tasks more passed
  2. #2 Qwen3.5 4B4.2B total parameters93.8%
    15 tasks more passed
  3. #3 Qwen3.5 0.8B Q4_00.8B total parameters37.5%
    6 tasks more passed
  4. #4 Gemma 4 E2B Q4_04.6B total parameters31.3%
    5 tasks more passed
  5. #5 Granite 4.0 Micro 3.4B Q4_03.4B total parameters25.0%
    4 tasks more passed
  6. #6 SmolLM3 3B Q4_K_M3.1B total parameters18.8%
    3 tasks more passed
  7. #7 LFM2.5 1.2B Q4_01.2B total parameters12.5%
    2 tasks more passed
  8. #8 Phi-4 Mini Instruct 3.8B Q4_K_M3.8B total parameters0.0%
    This model
How this comparison works

Share of tasks that passed the recorded checks. Family comparisons show the labeled Instruct and reasoning versions. Other peer comparisons use the same requested thinking setting, not necessarily each model’s best mode. Small differences on these short tests may not hold on other tasks. Peer membership varies by test coverage and compatible settings.

Comparison saved with the 2026-10-22 assessment. Task-specific results.

Logic and number puzzlesThinking off · Small · under 5B · dense

Solve newly generated puzzles using arithmetic, deduction and planning.

Mean puzzle credit · 0–100 points · higher is better
34 of 34 tasks scored for each model.

Gaps are relative to this model. Treat small gaps cautiously.

  1. #1 Qwen3.5 4B4.2B total parameters58.9/100
    32.9 points higher
  2. #2 Gemma 4 E2B Q4_04.6B total parameters58.9/100
    32.9 points higher
  3. #3 LFM2.5 1.2B Q4_01.2B total parameters38.3/100
    12.4 points higher
  4. #4 Qwen3.5 2B Q4_K_M1.9B total parameters35.4/100
    9.4 points higher
  5. #5 Granite 4.0 Micro 3.4B Q4_03.4B total parameters28.8/100
    2.8 points higher
  6. #6 Phi-4 Mini Reasoning 3.8B Q4_K_M3.8B total parameters26.5/100
    0.6 points higher
  7. #7 Phi-4 Mini Instruct 3.8B Q4_K_M3.8B total parameters26.0/100
    This model
  8. #8 SmolLM3 3B Q4_K_M3.1B total parameters15.2/100
    10.7 points lower
  9. #9 Qwen3.5 0.8B Q4_00.8B total parameters5.9/100
    20.0 points lower
How this comparison works

Average credit per task; partly correct results can earn partial credit. Family comparisons show the labeled Instruct and reasoning versions. Other peer comparisons use the same requested thinking setting, not necessarily each model’s best mode. Small differences on these short tests may not hold on other tasks. Peer membership varies by test coverage and compatible settings.

Comparison saved with the 2026-10-22 assessment. Task-specific results.

Published hard problemsThinking off · Small · under 5B · dense

Answer selected difficult math and reasoning questions from LiveBench.

Mean answer credit · 0–100 points · higher is better
11 of 11 tasks scored for each model.

Gaps are relative to this model. Treat small gaps cautiously.

  1. #1 Qwen3.5 4B4.2B total parameters55.8/100
    50.6 points higher
  2. #2 Gemma 4 E2B Q4_04.6B total parameters20.3/100
    15.1 points higher
  3. #3 Qwen3.5 2B Q4_K_M1.9B total parameters19.4/100
    14.2 points higher
  4. #4 SmolLM3 3B Q4_K_M3.1B total parameters14.9/100
    9.7 points higher
  5. #5 Granite 4.0 Micro 3.4B Q4_03.4B total parameters11.7/100
    6.5 points higher
  6. #6 Phi-4 Mini Reasoning 3.8B Q4_K_M3.8B total parameters10.5/100
    5.4 points higher
  7. #7 Phi-4 Mini Instruct 3.8B Q4_K_M3.8B total parameters5.2/100
    This model
  8. #8 LFM2.5 1.2B Q4_01.2B total parameters3.5/100
    1.7 points lower
  9. #9 Qwen3.5 0.8B Q4_00.8B total parameters3.3/100
    1.9 points lower
How this comparison works

Average credit per task; partly correct results can earn partial credit. Family comparisons show the labeled Instruct and reasoning versions. Other peer comparisons use the same requested thinking setting, not necessarily each model’s best mode. Small differences on these short tests may not hold on other tasks. Peer membership varies by test coverage and compatible settings.

Comparison saved with the 2026-10-22 assessment. Task-specific results.

Worth knowing

Thinking-on evidence is an incomplete, ineligible preview. The requested switch was ignored, so there is no verified thinking benefit to recommend for this configuration.

Evidence for this judgment

Source limitation
Some suite/condition results are incomplete or ineligible; their observed scores are previews.

Source limitation
Some requested reasoning conditions lack verified control.

Publication redactions limit what readers can inspect.

Evidence for this judgment

Source limitation
Some selected evidence has recorded publication redactions.

Tool-based results apply to the tested native-tool adapter. Successful diagnostic exchanges do not establish successful project work.

Evidence for this judgment

Source limitation
Agent tasks were released under a narrowly reviewed native-tool envelope adapter. Successful diagnostic tool exchanges do not imply reliable task performance: production task outcomes and unchanged graders govern the result. Earlier held and failed evidence remains in history. No repeated prompt tuning or scored rerolls were used.

How to interpret these assessments and their limitations →

One weight format is recorded for this model. Additional tested formats will appear here when available.

What changed in these results

Agent tasks were released under a narrowly reviewed native-tool envelope adapter. Successful diagnostic tool exchanges do not imply reliable task performance: production task outcomes and unchanged graders govern the result. Earlier held and failed evidence remains in history. No repeated prompt tuning or scored rerolls were used.

Thinking tradeoff

No verified on/off comparison is available for this configuration and test suite. Requested modes alone do not establish a controlled thinking comparison.

Sampled repetitions

Repeated runs are a separate experiment and never add votes to the baseline score.

Thinking off · temperature 0.6

48/48 repeats scored · 0 full-credit repeats.

16/16 questions have every planned repeat scored. Of those, 0 earned full credit on every repeat.

Recorded repetition conditions
  • Seed 303 · requested off · ignored switch
  • Seed 101 · requested off · ignored switch
  • Seed 202 · requested off · ignored switch

Everyday coding tasks: recorded outcomes

Find information in a small project, edit files and fix bugs.

16 incorrect

Each segment counts tasks, not credit points. Partial credit varies by task.

Scoring and fixed question bank →

This model’s answers, code and tool activity

Select a question to read the original answer, tool calls and available code changes for this configuration and thinking setting.

1–16 of 16 matching questions · 16 in this version

  • Locate the Retry Limit

    The user asks where the project's retry-limit setting is defined and what value it holds. This exercises locating a named setting in code and reporting both its location and value.

    basic-T1 · primary
  • Find the Release Date

    The user asks for the release date mentioned in a project notes document. This exercises reading a specified file and extracting a particular fact rather than guessing from outside knowledge.

    basic-T2 · primary
  • Change the Retry Limit

    The user asks to change the project's retry limit to a specified value. This exercises finding the relevant setting and making a focused code edit rather than merely describing the change.

    basic-T3 · primary
  • Create an Exact One-Line File

    The user asks to create a text file containing exactly one supplied line. This exercises basic file creation and precise instruction following, including avoiding extra text in the file.

    basic-T4 · primary
  • Explain the Latest Commit

    The user asks what the most recent Git commit changed in the project. This exercises inspecting recorded changes and turning them into a clear explanation of what was modified.

    basic-T5 · primary
  • Count and Compare Python Files

    The user asks how many Python files are in the application folder and which is largest. This exercises identifying relevant files, counting them, and comparing their sizes.

    basic-T6 · primary
  • Add Readable Byte Sizes

    The user asks to add a function that converts byte counts into readable sizes and remove the related TODO comment. This exercises implementing a small utility and updating the surrounding code consistently.

    basic-T7 · primary
  • Collect TODO Comments

    The user asks to find every TODO comment in the Python code and append each to the project's task list on a separate line. This combines searching across files with collecting text into an existing document.

    basic-T8 · primary
  • Fix Code Behind a Failing Test

    The user asks to investigate a failing test and fix the application code without changing the test. This exercises diagnosing a bug while respecting an explicit limit on what may be edited.

    basic-H1 · primary
  • Rename a Function Throughout

    The user asks to rename a network function and update every place in the project that uses it. This exercises tracking references across files and making a coordinated change rather than editing only the definition.

    basic-H2 · primary
  • Find the Proxy Settings

    The user asks to find a proxy configuration file somewhere under the project's parent folder and report its host and port. This exercises searching beyond the project folder and reading configuration values.

    basic-H3 · primary
  • Locate an Error in a Log

    The user asks which error code appears in a long event log and on which line. This exercises finding a specific entry in a large file and reporting its precise location.

    basic-H4 · primary
  • Append a Final Task

    The user asks to append a supplied release-related task as the final line of an existing task list. This exercises preserving existing content while adding exact text in the requested position.

    basic-H5 · primary
  • Calculate Total Retry Waiting

    The user asks for the total time a network function spends sleeping when all attempts fail, worked out from the code. This exercises following retry logic and adding the delays along that failure path.

    basic-H6 · primary
  • Document Every Command-Line Option

    The user asks for a project readme listing every command-line option with its help text. This exercises discovering the program's available options and transferring that information into complete, source-grounded documentation.

    basic-H7 · primary
  • Undo a Commit by Editing

    The user asks to undo the latest Git commit's change by editing the file directly, without using Git to make changes. This exercises understanding a previous edit and reversing it under a method constraint.

    basic-H8 · primary
Filters change this list, not the fixed scoring denominator. Search currently covers question IDs, titles and available task explanations. Focus a control to load prompt/outcome search where this edition provides it.

Suggested comparisons

These are starting points for inspection, not claims that one setup is better. Runtime and placement differences remain visible.

Qwen3.5 4B

Another checkpoint in the same total-parameter band and architecture category, with matching task controls.

Requested thinking off · Complete compared coverage

Compare these configurations →

Selection rule and evidence IDsPolicy comparisons-v3 · Template same-parameter-architecture-matching-controls
Facts: fact-b158bb60a71073ed7cff8c86, fact-f3caf8beeec8b9313dcb4eaf
Matching candidate set: fact-f3caf8beeec8b9313dcb4eaf, fact-c316deeab93502c3e19d46ba, fact-8cd25fbbd82bdae0d5893fac, fact-9e831223924ab330a13b6b40, fact-4bb7924a4c33187258857975, fact-f61609b3038ba8e6778d4e5a, fact-775ed46998b6dd1af2acd16c, fact-4c13720a63efcc49c96b4505

Tested weights

Q4_K_M · 2.49 GB

unsloth · recorded source

Phi-4-mini-instruct-Q4_K_M.gguf · Exact GGUF file →Historical first-MiB hash only; not a full-file identity.

Explore the answers

Choose a suite to inspect its exact available questions, references and model answers.