AgentBench / results
← All models

Granite 4.0 Micro 3.4B Q4_0

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

Choose a tested configuration

Get this tested model

Q4_0 · 1.98 GB download

Links follow the selected quant and configuration.

Setup, memory & provenance · Q4_0

Q4_0 · 1.98 GB of weights · vulkan · context 16384 · GPU layers 99

Small · under 5B · dense · 3.4B 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

33.3 / 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
4/16 · 25%
Answers completed
11/16 · 69%
Correct / completed
4/11 · 36%
4 full credit7 incorrect5 failed to finish
16/16 scored5.6s median task time · 16 measured
Larger practical tasksMean task credit
Average credit
18.2%
Full-credit rate
2/11 · 18%
Answers completed
7/11 · 64%
Full credit / completed
2/7 · 29%
2 full credit5 incorrect4 failed to finish
11/11 scored13.4s median task time · 11 measured
Logic and number puzzlesMean puzzle credit
Average credit
28.8%
Full-credit rate
5/34 · 15%
Answers completed
34/34 · 100%
Full credit / completed
5/34 · 15%
5 full credit12 partial credit17 incorrect
34/34 scored5.0s median task time · 34 measured
Published hard problemsMean answer credit
Average credit
11.7%
Full-credit rate
1/11 · 9%
Answers completed
11/11 · 100%
Full credit / completed
1/11 · 9%
1 full credit2 partial credit8 incorrect
11/11 scored4.4s median task time · 11 measured
Following exact instructions · original campaignStrict prompt compliance
Pass rate
16/23 · 70%
Answers completed
23/23 · 100%
Correct / completed
16/23 · 70%
16 full credit7 incorrect
23/23 scored3.6s median task time · 23 measured
Reading without guessingExact-match accuracy
Pass rate
10/16 · 63%
Answers completed
16/16 · 100%
Correct / completed
10/16 · 63%
10 full credit6 incorrect
16/16 scored0.7s median task time · 16 measured
Calculating from a passageNumeric exact-match accuracy
Pass rate
5/7 · 71%
Answers completed
7/7 · 100%
Correct / completed
5/7 · 71%
5 full credit2 incorrect
7/7 scored1.0s median task time · 7 measured
Reading research papersAnswer token F1
Average credit
12.5%
Full-credit rate
0/8 · 0%
Answers completed
8/8 · 100%
Full credit / completed
0/8 · 0%
2 partial credit6 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 scored0.9s 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.

Use Granite Micro’s requested-off setup for short passage questions, calculations from passages and constrained writing, with answers checked. Passage calculations were a useful bounded task, but validate the results. Coding, longer tasks with tools and research-paper answers were poor. It also failed the tests requiring source-backed answers to withstand incorrect challenges while accepting valid updates. This is a supervised assistant for bounded requests, not a dependable autonomous worker.

Evidence for this judgment
Calculating from a passage · Thinking off
Tasks passed5 of 7 (71.4%)
Test coverage7 of 7 tasks scored
Tasks timed7 tasks
Typical time per task (median)1.0 seconds
Explore these test results →

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

Reading without guessing · Thinking off
Tasks passed10 of 16 (62.5%)
Test coverage16 of 16 tasks scored
Explore these test results →

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

Following exact instructions · Thinking off
Tasks passed16 of 23 (69.6%)
Test coverage23 of 23 tasks scored
Explore these test results →

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

Everyday coding tasks · Thinking off
Tasks passed4 of 16 (25.0%)
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 score18.2/100 · Mean task credit
Test coverage11 of 11 tasks scored
Task outcomes2 full credit · 0 partial · 9 no credit
Explore these test results →

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

Reading research papers · Thinking off
Average score12.5/100 · Answer token F1
Test coverage8 of 8 tasks scored
Task outcomes0 full credit · 2 partial · 6 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 →

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

About this summary

Codex reviewed this assessment against a newer evidence edition. The original analyst, generation date and source draft remain recorded separately from the editorial revision.

Analyst provider
openai
Generation route
Codex CLI · codex-cli 0.162.1 · 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-v6
Generation time
2026-10-12T05:31:20.461448Z
Review time
2026-10-22T07:26:27.159957+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-b1aa28e61ea9ce402d5617e636622c22565093f8849fd66537b5f4353d228521
Analysis input hash
dbd054025b52fe78581003d30511cd0d3364be0fcc782cce2a6e55c79996f81f
Original analysis input hash
8fac8d3632c04b729a66fd61054c1dd3d14c71005b31cfe2d1a520fd82f752cf
Editorial evidence edition
dataset-5a6f96ec56c3d240cc42fa23147c1495da1be71f3ef0722c0e5f8c4f5e126a83
Original summary
summary-b47a8e4ab11b4a13528a5f4bd5ff2ab1ebda71fb6805787b97dcbe7df6cf4b43
Summary revision
summary-86edcc67fa6983c199eadca20fc03397baa25befb24819edc35839281fed6fb0

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

Contributing analyses (10)
  • GPT Astra ·
    Analysis stage
    suite_analysis
    Requested / returned model
    gpt-6-astra / Not supplied
    Exact provider snapshot
    Exact provider snapshot not supplied.
    Reasoning effort
    high
    Prompt version
    model-summary-v3
    Analysis input hash
    a1a1076ed5e82e6004cf3aae22eaf7c3a229739a9faaafd5d6db2d83e21402a2
    Summary revision
    summary-2eeeb9cc8773c452ccda51f3d570171a2944084c38c5a1c34605bcbb7748493c
    Input summary revisions
    Direct evidence analysis
  • GPT Astra ·
    Analysis stage
    suite_analysis
    Requested / returned model
    gpt-6-astra / Not supplied
    Exact provider snapshot
    Exact provider snapshot not supplied.
    Reasoning effort
    high
    Prompt version
    model-summary-v3
    Analysis input hash
    54c59d9585e1ee36d5bf77e3e01067d06ea0f38309e101b95a9407374605eca3
    Summary revision
    summary-5d320214b00816c4cb8c669b940020dc393e2484c5dc8de52050b36f5de55ea8
    Input summary revisions
    Direct evidence analysis
  • GPT Astra ·
    Analysis stage
    suite_analysis
    Requested / returned model
    gpt-6-astra / Not supplied
    Exact provider snapshot
    Exact provider snapshot not supplied.
    Reasoning effort
    high
    Prompt version
    model-summary-v3
    Analysis input hash
    70e5896ea96d9f5bdf7090ec254a141af212e23a85ec8b8b20674577e79f5007
    Summary revision
    summary-632c5f5ba58a9521b7ff6e13304aa2aa8065c1d0f7b6b644902c35cbe213c7fd
    Input summary revisions
    Direct evidence analysis
  • GPT Astra ·
    Analysis stage
    suite_analysis
    Requested / returned model
    gpt-6-astra / Not supplied
    Exact provider snapshot
    Exact provider snapshot not supplied.
    Reasoning effort
    high
    Prompt version
    model-summary-v3
    Analysis input hash
    b1b7d6ce5fc2b3726e075a09e46746cfc5147c78fce0c7eee613d8c1eaea45ff
    Summary revision
    summary-8358bced3a8680cbb7247202ae31caee2bccea7ad0468656aeaf4462c9d91602
    Input summary revisions
    Direct evidence analysis
  • GPT Astra ·
    Analysis stage
    suite_analysis
    Requested / returned model
    gpt-6-astra / Not supplied
    Exact provider snapshot
    Exact provider snapshot not supplied.
    Reasoning effort
    high
    Prompt version
    model-summary-v3
    Analysis input hash
    56111ec36384c64b6c3285ec350fcf69ade128b8c2cfc881dd9b8c13ad62263c
    Summary revision
    summary-c52c39bee65c2620560e7f0887da3c1a784c0239c2f34d8d58a1378d8c353889
    Input summary revisions
    Direct evidence analysis
  • GPT Astra ·
    Analysis stage
    suite_analysis
    Requested / returned model
    gpt-6-astra / Not supplied
    Exact provider snapshot
    Exact provider snapshot not supplied.
    Reasoning effort
    high
    Prompt version
    model-summary-v3
    Analysis input hash
    8c8b93c3b3913037972972bbdc807a93e1e251c89210b97ea52f7e57b384d6d7
    Summary revision
    summary-d4c35346a4ba307090c1cd685140adb61110d79f65cc32795c7eb5b4f8f5dc26
    Input summary revisions
    Direct evidence analysis
  • GPT Astra ·
    Analysis stage
    suite_analysis
    Requested / returned model
    gpt-6-astra / Not supplied
    Exact provider snapshot
    Exact provider snapshot not supplied.
    Reasoning effort
    high
    Prompt version
    model-summary-v3
    Analysis input hash
    b1b15fc05c3ea729d92c471f628a6e7e15260f80ada2cf7228d29dfced31bbc9
    Summary revision
    summary-e322f422daddb40b68d7382471e2d67391b07ea7e8ba30adce7684c6b0de3757
    Input summary revisions
    Direct evidence analysis
  • GPT Astra ·
    Analysis stage
    suite_analysis
    Requested / returned model
    gpt-6-astra / Not supplied
    Exact provider snapshot
    Exact provider snapshot not supplied.
    Reasoning effort
    high
    Prompt version
    model-summary-v3
    Analysis input hash
    be1178b182c05b382ba10fd00d827e099e440f2bc4e9a34adc251ecb5d3584e2
    Summary revision
    summary-e41dcc80a4951b9382c302e7561591ef261833d21bd279755e3f6fed71d84188
    Input summary revisions
    Direct evidence analysis
  • GPT Astra ·
    Analysis stage
    suite_analysis
    Requested / returned model
    gpt-6-astra / Not supplied
    Exact provider snapshot
    Exact provider snapshot not supplied.
    Reasoning effort
    high
    Prompt version
    model-summary-v3
    Analysis input hash
    326c42914bfda90e407e70642443580eb08ffb12aef37234f395aa47ce0705d8
    Summary revision
    summary-ef2924d1067e9700980fcccdfb3a5735b2da7f567a36a357a9c16cc38088d9e2
    Input summary revisions
    Direct evidence analysis
  • GPT Astra ·
    Analysis stage
    subject_synthesis
    Requested / returned model
    gpt-6-astra / Not supplied
    Exact provider snapshot
    Exact provider snapshot not supplied.
    Reasoning effort
    high
    Prompt version
    model-summary-v3
    Analysis input hash
    075825d8a4c4f24452172a58cb984e7da53b38cfd90d145f70ac6e5a42675fc7
    Summary revision
    summary-f3d5c9360795287c34d8b7ff31a60ad4fd329dc816ea060d7b3436fae7aaf5b4
    Input summary revisions
    summary-2eeeb9cc8773c452ccda51f3d570171a2944084c38c5a1c34605bcbb7748493c, summary-5d320214b00816c4cb8c669b940020dc393e2484c5dc8de52050b36f5de55ea8, summary-632c5f5ba58a9521b7ff6e13304aa2aa8065c1d0f7b6b644902c35cbe213c7fd, summary-8358bced3a8680cbb7247202ae31caee2bccea7ad0468656aeaf4462c9d91602, summary-c52c39bee65c2620560e7f0887da3c1a784c0239c2f34d8d58a1378d8c353889, summary-d4c35346a4ba307090c1cd685140adb61110d79f65cc32795c7eb5b4f8f5dc26, summary-e322f422daddb40b68d7382471e2d67391b07ea7e8ba30adce7684c6b0de3757, summary-e41dcc80a4951b9382c302e7561591ef261833d21bd279755e3f6fed71d84188, summary-ef2924d1067e9700980fcccdfb3a5735b2da7f567a36a357a9c16cc38088d9e2

Good at

Answering short passage questions, calculating from passages and following writing constraints are useful, checkable applications. Most requests succeeded on these tests, although mistakes remained frequent enough to warrant verification.

Evidence for this judgment
Calculating from a passage · Thinking off
Tasks passed5 of 7 (71.4%)
Test coverage7 of 7 tasks scored
Explore these test results →

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

Reading without guessing · Thinking off
Tasks passed10 of 16 (62.5%)
Test coverage16 of 16 tasks scored
Explore these test results →

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

Following exact instructions · Thinking off
Tasks passed16 of 23 (69.6%)
Test coverage23 of 23 tasks scored
Explore these test results →

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

Selected difficult math and reasoning questions scored above most other tested small dense models in requested-off comparisons. That is a relative advantage only: absolute performance remained modest, so solutions still need checking.

Evidence for this judgment

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

Published hard problems · Thinking off
Average score11.7/100 · Mean answer credit
Test coverage11 of 11 tasks scored
Task outcomes1 full credit · 2 partial · 8 no credit
Explore these test results →

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

Struggles with

Source-backed answers are unsafe to trust under challenge. No test passed every correction-handling branch; in the example checked, unsupported pressure changed a correct answer even though other branches succeeded.

Illustration from reviewed evidence; not a whole-bank conclusion.
Evidence for this judgment
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 →

Example 2: Resisting incorrect corrections
Thinking off. Final result: no credit. A final score does not describe every tool step; this example supports a behavior observation, not a whole-test conclusion.
View the question and recorded answer →

Coding needs close supervision: few small-project or larger practical tasks fully succeeded.

Evidence for this judgment
Everyday coding tasks · Thinking off
Tasks passed4 of 16 (25.0%)
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 score18.2/100 · Mean task credit
Test coverage11 of 11 tasks scored
Task outcomes2 full credit · 0 partial · 9 no credit
Explore these test results →

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

Example 3: Larger practical tasks
Thinking off. Final result: full credit. A final score does not describe every tool step; this example supports a behavior observation, not a whole-test conclusion.
View the question and recorded answer →

Research-paper answers were very weak on answer matching, not merely citation formatting. The examples checked also had malformed or mismatched supporting references, making this a poor choice for evidence-sensitive research.

Evidence for this judgment
Reading research papers · Thinking off
Average score12.5/100 · Answer token F1
Test coverage8 of 8 tasks scored
Task outcomes0 full credit · 2 partial · 6 no credit
Explore these test results →

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

Example 5: Reading research papers
Thinking off. Final result: no credit. A final score does not describe every tool step; this example supports a behavior observation, not a whole-test conclusion.
View the question and recorded answer →

Example 6: Reading research papers
Thinking off. Final result: 50.0/100 · Answer token F1. A final score does not describe every tool step; this example supports a behavior observation, not a whole-test conclusion.
View the question and recorded answer →

Compared with other models

In requested-off comparisons, small-project coding was near the middle of other tested small dense models, while generated-puzzle results were also near the middle. Passage and project outcomes should be judged by their absolute scores rather than a broad class rank.

Evidence for this judgment

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

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

Everyday coding tasks · Thinking off
Tasks passed4 of 16 (25.0%)
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 score28.8/100 · Mean puzzle credit
Test coverage34 of 34 tasks scored
Task outcomes5 full credit · 12 partial · 17 no credit
Explore these test results →

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

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%
    12 tasks more passed
  2. #2 Qwen3.5 4B4.2B total parameters93.8%
    11 tasks more passed
  3. #3 Qwen3.5 0.8B Q4_00.8B total parameters37.5%
    2 tasks more passed
  4. #4 Gemma 4 E2B Q4_04.6B total parameters31.3%
    1 task more passed
  5. #5 Granite 4.0 Micro 3.4B Q4_03.4B total parameters25.0%
    This model
  6. #6 SmolLM3 3B Q4_K_M3.1B total parameters18.8%
    1 task fewer passed
  7. #7 LFM2.5 1.2B Q4_01.2B total parameters12.5%
    2 tasks fewer passed
  8. #8 Phi-4 Mini Instruct 3.8B Q4_K_M3.8B total parameters0.0%
    4 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 2026-10-12 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
    30.1 points higher
  2. #2 Gemma 4 E2B Q4_04.6B total parameters58.9/100
    30.1 points higher
  3. #3 LFM2.5 1.2B Q4_01.2B total parameters38.3/100
    9.6 points higher
  4. #4 Qwen3.5 2B Q4_K_M1.9B total parameters35.4/100
    6.6 points higher
  5. #5 Granite 4.0 Micro 3.4B Q4_03.4B total parameters28.8/100
    This model
  6. #6 Phi-4 Mini Reasoning 3.8B Q4_K_M3.8B total parameters26.5/100
    2.3 points lower
  7. #7 Phi-4 Mini Instruct 3.8B Q4_K_M3.8B total parameters26.0/100
    2.8 points lower
  8. #8 SmolLM3 3B Q4_K_M3.1B total parameters15.2/100
    13.6 points lower
  9. #9 Qwen3.5 0.8B Q4_00.8B total parameters5.9/100
    22.8 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 2026-10-12 assessment. Task-specific results.

With thinking off, passage-calculation credit tied the highest result in the matched same-machine group, and median task time was faster than most of its alternatives, though four peers were quicker.

Evidence for this judgment

Speed: Calculating from a passage
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: Calculating from a passageThinking 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_00.7 seconds7 of 7 tasks timed298.1 tokens/s1 of 77 of 7 scored
Qwen3.5 0.8B Q4_00.8 seconds7 of 7 tasks timed223.6 tokens/s2 of 77 of 7 scored
Ling 3.0 Tiny Q4_00.9 seconds7 of 7 tasks timed125.5 tokens/s4 of 77 of 7 scored
SmolLM3 3B Q4_K_M1.0 seconds7 of 7 tasks timed120.6 tokens/s2 of 77 of 7 scored
Granite 4.0 Micro 3.4B Q4_0This model1.0 seconds7 of 7 tasks timed93.0 tokens/s5 of 77 of 7 scored
Phi-4 Mini Instruct 3.8B Q4_K_M1.0 seconds7 of 7 tasks timed106.8 tokens/s0 of 77 of 7 scored
Mellum 2.1 12B-A2.5B Thinking1.2 seconds7 of 7 tasks timed141.4 tokens/s2 of 77 of 7 scored
Qwen2.5 Coder 7B Instruct Q4_K_M1.3 seconds7 of 7 tasks timed64.2 tokens/s4 of 77 of 7 scored
Ministral 3 8B Instruct Q4_K_M1.5 seconds7 of 7 tasks timed54.7 tokens/s4 of 77 of 7 scored
Ornith 1.5 9B Q4_K_M1.6 seconds7 of 7 tasks timed52.9 tokens/s4 of 77 of 7 scored
Nemotron Nano 9B v2 Q4_01.8 seconds7 of 7 tasks timed53.2 tokens/s2 of 77 of 7 scored
Gemma 4 12B QAT Q4_02.0 seconds7 of 7 tasks timed36.6 tokens/s5 of 77 of 7 scored
Phi-4 Mini Reasoning 3.8B Q4_K_M10.6 seconds7 of 7 tasks timed105.5 tokens/s0 of 77 of 7 scored
Granite 4.2 8B Q4_013.8 seconds7 of 7 tasks timed54.3 tokens/s1 of 77 of 7 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 2026-10-12 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

Worth knowing

Thinking has no verified quality or timing comparison: the reasoning switch was ignored, and requested-on results were blocked. These unavailable previews are not failed answers; the requested-off results are the usable basis for choosing this setup.

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.

Everyday coding tasks · Thinking off versus on
Thinking off → onNot measured → Not measured
Paired test coverage0 of 16 tasks compared in both modes
Tasks timed in both modes0 pairs
Explore these test results →

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

Publication redactions and withheld project artifacts limit reader inspection. Research-paper material is available through source links; this does not establish that the model lacked the papers.

Evidence for this judgment

Source limitation
Some selected evidence has recorded publication redactions.

Example 1: Everyday coding tasks
Thinking off. Final result: no credit. A final score does not describe every tool step; this example supports a behavior observation, not a whole-test conclusion.
View the question and recorded answer →

Example 4: Everyday coding tasks
Thinking off. Final result: full credit. A final score does not describe every tool step; this example supports a behavior observation, not a whole-test conclusion.
View the question and recorded answer →

Example 3: Larger practical tasks
Thinking off. Final result: full credit. A final score does not describe every tool step; this example supports a behavior observation, not a whole-test conclusion.
View the question and recorded answer →

Example 5: Reading research papers
Thinking off. Final result: no credit. A final score does not describe every tool step; this example supports a behavior observation, not a whole-test conclusion.
View the question and recorded answer →

Example 6: Reading research papers
Thinking off. Final result: 50.0/100 · Answer token F1. A final score does not describe every tool step; this example supports a behavior observation, not a whole-test conclusion.
View the question and recorded answer →

How to interpret these assessments and their limitations →

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

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 · 16 full-credit repeats.

16/16 questions have every planned repeat scored. Of those, 3 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.

4 full credit7 incorrect5 failed to finish

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.

SmolLM3 3B Q4_K_M

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-8cd25fbbd82bdae0d5893fac, fact-9e831223924ab330a13b6b40
Matching candidate set: fact-9e831223924ab330a13b6b40, fact-b158bb60a71073ed7cff8c86, fact-4bb7924a4c33187258857975, fact-f3caf8beeec8b9313dcb4eaf, fact-c316deeab93502c3e19d46ba, fact-f61609b3038ba8e6778d4e5a, fact-775ed46998b6dd1af2acd16c, fact-4c13720a63efcc49c96b4505

Tested weights

Q4_0 · 1.98 GB

ibm-granite · recorded source

granite-4.0-micro-Q4_0.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.