Benchmark evidence status

Benchmark evidence28 August 2026

Where the benchmark evidence stands

No same-task comparison has passed the scorecard gate yet, so none of the 6 named baseline tools has a result; the 2 measured comparisons are published separately with their own limits.

  • 0

    same-task comparison groups passed the scorecard gate

  • 2

    measured comparisons published on their own page

  • 6

    named baseline tools not measured yet

    Aider, Claude Code, Cline, OpenAI Codex, OpenCode and OpenHands

  • 112

    benchmark-like records kept out of any comparison

Read the measured results: The seven-case stack matrix · The 164-task model comparison

A tool that has not been measured has no score here, which is not a score of zero.

How we know: the evidence status
Does not prove
A same-task-set comparison does not establish general model quality, production reliability, causal harness advantage, security, adoption, or independent validation. Results remain scoped to the recorded tasks, prompts, policy, models, harnesses, commit, and metric schema.
Scorecard gate
Generic scorecard gate: unknown · no_valid_comparable_scorecards_found. 0 same-task-set comparison groups passed that gate. 2 separately scoped, source-hashed actual comparisons are published with their own limitations. A target marked NOT_MEASURED is a planned baseline, not a result and not a zero.
Named baseline targets
Market-adjacent and focal tools named by the site-owned benchmark plan.
ToolModelStatusReasonPlan provenance
Aidercaptured at executionNOT_MEASUREDNo same-task, same-environment executed scorecard is present in the public analytics capture.official · source
Claude Codecaptured at executionNOT_MEASUREDNo same-task, same-environment executed scorecard is present in the public analytics capture.official · source
Clinecaptured at executionNOT_MEASUREDNo same-task, same-environment executed scorecard is present in the public analytics capture.official · source
OpenAI Codexcaptured at executionNOT_MEASUREDNo same-task, same-environment executed scorecard is present in the public analytics capture.official · source
OpenCodecaptured at executionNOT_MEASUREDNo same-task, same-environment executed scorecard is present in the public analytics capture.official · source
OpenHandscaptured at executionNOT_MEASUREDNo same-task, same-environment executed scorecard is present in the public analytics capture.official · source
Excluded benchmark candidates
Tracked benchmark-like JSON kept out of generic comparative figures.
ProjectTracked pathSchemaExclusion reason
flywheelartifacts/agent_recovery_benchmark/20260708_225412/agent_recovery_benchmark.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_225850/agent_recovery_benchmark.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_230039/agent_recovery_benchmark.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_230336/agent_recovery_benchmark.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_230513/agent_recovery_benchmark.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_230634/agent_recovery_benchmark.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_230923/spin/artifacts/agent_recovery_benchmark/20260708_230923/artifacts/agent_recovery_benchmark/20260708_230923/forum_benchmark_cases.jsonforum.benchmark-case-set/v1unsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_230923/spin/artifacts/agent_recovery_benchmark/20260708_230923/artifacts/agent_recovery_benchmark/20260708_230923/loop/cache/e7d325c9cd42556a.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_230923/spin/artifacts/agent_recovery_benchmark/20260708_230923/artifacts/agent_recovery_benchmark/20260708_230923/loop/env/example_pass-96a48c08fee968f6.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_230923/spin/artifacts/agent_recovery_benchmark/20260708_230923/artifacts/agent_recovery_benchmark/20260708_230923/m7_flywheel_local_dry.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheelartifacts/agent_recovery_benchmark/20260708_230923/spin/artifacts/agent_recovery_benchmark/20260708_230923/artifacts/agent_recovery_benchmark/20260708_230923/m7_frontier_codex_dry.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheelartifacts/agent_recovery_benchmark/20260708_230923/spin/artifacts/agent_recovery_benchmark/20260708_230923/artifacts/agent_recovery_benchmark/20260708_230923/report.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_230923/spin/fw_fail_a/task.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_230923/spin/fw_fail_b/task.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_230923/spin/fw_pass_a/task.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_230923/spin/fw_pass_b/task.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_231022/forum_benchmark_cases.jsonforum.benchmark-case-set/v1unsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_231022/loop/cache/e7d325c9cd42556a.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_231022/loop/env/example_pass-96a48c08fee968f6.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_231022/m7_flywheel_local_dry.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheelartifacts/agent_recovery_benchmark/20260708_231022/m7_frontier_codex_dry.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheelartifacts/agent_recovery_benchmark/20260708_231022/report.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_231022/spin/fw_fail_a/task.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_231022/spin/fw_fail_b/task.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_231022/spin/fw_pass_a/task.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_231022/spin/fw_pass_b/task.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_231654/forum_benchmark_cases.jsonforum.benchmark-case-set/v1unsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_231654/loop/cache/e7d325c9cd42556a.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_231654/loop/env/example_pass-96a48c08fee968f6.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_231654/m7_flywheel_local_dry.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheelartifacts/agent_recovery_benchmark/20260708_231654/m7_frontier_codex_dry.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheelartifacts/agent_recovery_benchmark/20260708_231654/report.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_231654/spin/fw_fail_a/task.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_231654/spin/fw_fail_b/task.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_231654/spin/fw_pass_a/task.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_231654/spin/fw_pass_b/task.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_232208/forum_benchmark_cases.jsonforum.benchmark-case-set/v1unsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_232208/loop/cache/e7d325c9cd42556a.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_232208/loop/env/example_pass-96a48c08fee968f6.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_232208/m7_flywheel_local_dry.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheelartifacts/agent_recovery_benchmark/20260708_232208/m7_frontier_codex_dry.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheelartifacts/agent_recovery_benchmark/20260708_232208/report.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_232208/spin/fw_fail_a/task.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_232208/spin/fw_fail_b/task.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_232208/spin/fw_pass_a/task.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_232208/spin/fw_pass_b/task.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_232512/forum_benchmark_cases.jsonforum.benchmark-case-set/v1unsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_232512/loop/cache/e7d325c9cd42556a.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_232512/loop/env/example_pass-96a48c08fee968f6.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_232512/m7_flywheel_local_dry.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheelartifacts/agent_recovery_benchmark/20260708_232512/m7_frontier_codex_dry.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheelartifacts/agent_recovery_benchmark/20260708_232512/report.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_232512/spin/fw_fail_a/task.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_232512/spin/fw_fail_b/task.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_232512/spin/fw_pass_a/task.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_232512/spin/fw_pass_b/task.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_232931/forum_benchmark_cases.jsonforum.benchmark-case-set/v1unsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_232931/loop/cache/e7d325c9cd42556a.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_232931/loop/env/example_pass-96a48c08fee968f6.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_232931/m7_flywheel_local_dry.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheelartifacts/agent_recovery_benchmark/20260708_232931/m7_frontier_codex_dry.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheelartifacts/agent_recovery_benchmark/20260708_232931/report.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_232931/spin/fw_fail_a/task.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_232931/spin/fw_fail_b/task.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_232931/spin/fw_pass_a/task.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/agent_recovery_benchmark/20260708_232931/spin/fw_pass_b/task.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/flywheel-local-coder-14b-benchmark-ci.jsonunknownunsupported_benchmark_result_schema
flywheelartifacts/flywheel-local-coder-14b-benchmark-m7-hard-scorecard.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheelartifacts/flywheel-local-coder-14b-benchmark-m7-scorecard.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheelartifacts_codex_dry.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheelartifacts_codex_dry_easy_1task.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheelartifacts_codex_dry_easy_full.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheelartifacts_codex_dry_full_attempt.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheelartifacts_codex_frontier_dry_easy_full.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheelartifacts_codex_frontier_dry_hard_full.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheelartifacts_codex_harness_dry_easy.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheelartifacts_codex_live.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheelartifacts_local_ollama_easy_full.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheelartifacts_local_ollama_hard_full.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheelartifacts_local_run_2.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheelbenchmarks/agentic-task-set-adapter-v1.jsonharness.agentic-task-set-adapter/v1planning_or_contract_artifact_not_results
flywheelbenchmarks/agentic-task-set-implementation-plan-v1.jsonharness.agentic-task-set-implementation-plan/v1planning_or_contract_artifact_not_results
flywheelbenchmarks/agentic-task-set-v1.jsonharness.agentic-task-set/v1unsupported_benchmark_result_schema
flywheelbenchmarks/cross-harness-adapter-contract-v1.jsonharness.cross-harness-adapter-contract/v1planning_or_contract_artifact_not_results
flywheelbenchmarks/cross-harness-adapter-contract-v2.jsonharness.cross-harness-adapter-contract/v2planning_or_contract_artifact_not_results
flywheelbenchmarks/dry-run-preflight-command-deck-v1.jsonharness.dry-run-preflight-command-deck/v1unsupported_benchmark_result_schema
flywheelbenchmarks/embodied-realtime-multimodal-v1.jsonharness.benchmark-contract/v1planning_or_contract_artifact_not_results
flywheelbenchmarks/fixtures/cross-harness/documentation-topology-v1.jsonharness.cross-harness-fixture.documentation-topology/v1unsupported_benchmark_result_schema
flywheelbenchmarks/fixtures/cross-harness/index-events-v1.jsonharness.cross-harness-fixture.index-events/v1unsupported_benchmark_result_schema
flywheelbenchmarks/fixtures/cross-harness/paired-friction-observations-v1.jsonharness.cross-harness-fixture.paired-friction-observations/v1unsupported_benchmark_result_schema
flywheelbenchmarks/fixtures/cross-harness/shared-task-facts-v1.jsonharness.cross-harness-fixture.shared-task-facts/v1unsupported_benchmark_result_schema
flywheelbenchmarks/fixtures/evidence-journey/durable-restart-v2.jsonflywheel.evidence-journey-acceptance-fixture/v2unsupported_benchmark_result_schema
flywheelbenchmarks/forum-ledger-deep-verify-scaling-v1.jsonharness.benchmark-contract/v1planning_or_contract_artifact_not_results
flywheelhandoff/site-designer/evidence/benchmark-ci.jsonunknownunsupported_benchmark_result_schema
flywheelm7_scorecard.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheelproject-docs/releases/14B/shipped-page/benchmarks/easy_set_8task_scorecard.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheelproject-docs/releases/14B/shipped-page/benchmarks/hard_set_scorecard_QUARANTINED.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheelproject-docs/releases/14B/shipped-page/benchmarks/he_base_comparison.jsonflywheel.he-base-comparison/v1unsupported_benchmark_result_schema
flywheelproject-docs/releases/14B/shipped-page/benchmarks/humaneval_base_qwen14b.jsonunknownunsupported_benchmark_result_schema
flywheelproject-docs/releases/14B/shipped-page/benchmarks/humaneval_flywheel14b.jsonunknownunsupported_benchmark_result_schema
flywheeltasks/research/encoding_sweep_scorecard_20260706.jsonunknownunsupported_benchmark_result_schema
flywheeltasks/research/flexible_marker_scorecard_20260706.jsonunknownunsupported_benchmark_result_schema
flywheeltasks/research/m7_hard_scorecard_20260706.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheeltasks/research/m7_scorecard_20260706.jsonm7-scorecard/v1legacy_summary_missing_task_set_and_execution_identity
flywheeltasks/research/perception_scorecard_20260706.jsonunknownunsupported_benchmark_result_schema
forumbenchmarks/otlp_metrics_sample.jsonunknownunsupported_benchmark_result_schema
forumbenchmarks/otlp_trace_sample.jsonunknownunsupported_benchmark_result_schema
forumbenchmarks/storage_scaling.jsonforum.storage-scaling-benchmark/v1unsupported_benchmark_result_schema
telosdocs/outreach/receipts/twenty-second-wave/hyphal-context-benchmark-2026-07-02.jsonproject-telos.hyphal-context-benchmark/v1unsupported_benchmark_result_schema
telosdocs/outreach/receipts/twenty-second-wave/hyphal-context-benchmark-correction-2026-07-18.jsonproject-telos.hyphal-context-benchmark-correction/v1unsupported_benchmark_result_schema
telosdocs/outreach/receipts/twenty-second-wave/hyphal-context-benchmark.learn-packet.jsonunknownunsupported_benchmark_result_schema
telosdocs/outreach/receipts/twenty-second-wave/learn-hyphal-context/tutor/twenty-second-wave-hyphal-context-benchmark.prooflesson.jsonunknownunsupported_benchmark_result_schema

Status captured 28 August 2026