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: inconclusive · 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
20.6 / 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.
This analysis covers this configuration across its recorded thinking conditions and test suites.
Consider the tested requested-off setup for quick, low-stakes questions about a passage, with answers checked. It handles that work reasonably well, but it is a poor choice for unattended coding or longer tool-based work. Tests expose serious weaknesses in keeping answers source-backed when challenged or evidence changes. This is a narrow lookup option, not a general-purpose recommendation.
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
2026-10-12T03:52:21.338864Z
Review time
2026-10-22T07:26:27.211618+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.
Finding answers in passages is a usable, limited niche for low-stakes lookup. Most answers matched, though enough were wrong to make checking important.
Small project tasks succeeded only in a minority of cases; longer practical tasks rarely reached completion. Expect substantial supervision rather than handing it a project and trusting the result.
Do not rely on it to preserve facts under challenge and update them correctly: no source-correction scenario met every requirement. That does not mean every individual response was wrong.
Difficult math and reasoning, puzzles, and research-paper questions produced extremely low scores—not merely formatting slips. Strict wording and formatting instructions also failed more often than they succeeded.
With thinking requested off, passage answering ties the best other tested small dense models, despite substantial mistakes. Strict instruction following and puzzle scores place it near the bottom of that class; it ties one eligible peer on strict instruction following. On passage questions with thinking off, it ranked near the fastest in the matched same-machine comparison; the single faster alternative had a much lower answer score.
Evidence for this judgment
Reading without guessing Thinking off. Small · under 5B · dense. Tied at rank 1 of 9 in this recorded comparison.
Following exact instructions Thinking off. Small · under 5B · dense. Tied at rank 8 of 9 in this recorded comparison.
Logic and number puzzles Thinking off. Small · under 5B · dense. Rank 9 of 9 in this recorded comparison.
Speed: Reading without guessing Thinking off. Matching test-machine label and execution settings; spans model sizes. Timings include unsuccessful tasks, so read them alongside quality.
Reading without guessingThinking off · Small · under 5B · dense
Answer from a passage—or say that the passage does not contain the answer.
Exact-match accuracy · 0–100% · higher is better 16 of 16 tasks scored for each model.
Gaps are relative to this model. Treat small gaps cautiously.
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.
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.
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.
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.
Speed and task success: Reading without guessingThinking 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.
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
Some requested conditions are incomplete, ineligible, or lack verified reasoning control. Their scores are previews; requested-on results do not establish a thinking benefit or cost.
Evidence for this judgment
Source limitation Additional reasoning-control diagnostics remain inconclusive. Requested-off output is not proof that reasoning was suppressed; no matched thinking benefit or cost is established.
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 restrictions limit reader inspection of selected paper inputs and project artifacts. Omitted content cannot support additional conclusions.
Evidence for this judgment
Source limitation Some selected evidence has recorded publication redactions.
Example 2: Reading research papers Thinking off. Final result: 33.3/100 · 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 →
Example 3: 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 1: Larger practical 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: 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 →
One same-budget retry after context exhaustion. A second context limit is failed to answer (zero task-completion credit), not an incorrect submitted answer. Answer-only averages exclude these no-answer outcomes; completed retry answers enter the normal scoring path. This follow-up is selected after observing failures.
Evidence for this judgment
Source limitation One same-budget retry after context exhaustion. A second context limit is failed to answer (zero task-completion credit), not an incorrect submitted answer. Answer-only averages exclude these no-answer outcomes; completed retry answers enter the normal scoring path. This follow-up is selected after observing failures.
One weight format is recorded for this model. Additional tested formats will appear here when available.
What changed in these results
Additional reasoning-control diagnostics remain inconclusive. Requested-off output is not proof that reasoning was suppressed; no matched thinking benefit or cost is established.
One same-budget retry after context exhaustion. A second context limit is failed to answer (zero task-completion credit), not an incorrect submitted answer. Answer-only averages exclude these no-answer outcomes; completed retry answers enter the normal scoring path. This follow-up is selected after observing failures.
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 · 11 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 · inconclusive
Seed 202 · requested off · inconclusive
Seed 101 · requested off · inconclusive
Everyday coding tasks: recorded outcomes
Find information in a small project, edit files and fix bugs.
6 full credit4 incorrect6 failed to finish
Each segment counts tasks, not credit points. Partial credit varies by task.
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.
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.
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.
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.
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.
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.
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.
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.
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.
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.
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.
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.
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.
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.
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.
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.
LFM2.5 1.2B Q4_0
Another checkpoint in the same total-parameter band and architecture category, with matching task controls.
Requested thinking off · Complete compared coverage