GPT-5.5 Codexの推論トークンクラスタリングがパフォーマンス低下を招いている可能性

Uh oh! There was an error while loading. Please reload this page. openai / codex Public Notifications You must be signed in to change notification settings Fork 14.2k Star 95.5k GPT-5.5 Codex reasoning-token clustering at 516/1034/1552 may be leading to degraded performance on complex tasks #30364 New issueCopy link New issueCopy link OpenOpen GPT-5.5 Codex reasoning-token clustering at 516/1034/1552 may be leading to degraded performance on complex tasks#30364Copy link Labels bug Something isn't working Something isn't working model-behavior Issues related to behaviors exhibited by the model Issues related to behaviors exhibited by the model rate-limits Issues related to rate limits, quotas, and token usage reporting Issues related to rate limits, quotas, and token usage reporting Description vguptaa45 opened on Jun 27, 2026 Issue body actions Summary I found an aggregate pattern in Codex token_count metadata: gpt-5.5 responses disproportionately land at exactly reasoning_output_tokens = 516, with additional fixed-boundary spikes around 1034 and 1552. This appears model-specific and coincides with lower overall reasoning-token intensity, which may help explain degraded performance on complex/high-stakes Codex tasks. This is related to #29353, which reported a task-level reproduction where gpt-5.5 runs ending at exactly 516 reasoning tokens returned the wrong answer. This issue adds aggregate evidence across a larger Feb-Jun window. I am not claiming this proves hidden chain-of-thought truncation. The narrower claim is that Codex telemetry shows a GPT-5.5-specific fixed-token clustering anomaly that looks consistent with thresholded reasoning-budget behavior. Environment Product: Codex Model most implicated: gpt-5.5 Data source: Codex token_count metadata Time window analyzed: Feb 1-Jun 27, 2026 UTC Related issue: gpt-5.5 xhigh sometimes short-circuits with reasoning_output_tokens=516 and wrong final_answer in Codex Desktop #29353 Evidence Metric Value Response-level token records analyzed 390,195 Sessions represented 865 Exact reasoning_output_tokens = 516 events 3,363 GPT-5.5 share of all responses 19.3% GPT-5.5 share of exact-516 events 82.0% GPT-5.5 exact-516 / >=516 ratio 44.0% Non-GPT-5.5 exact-516 / >=516 ratio 1.3% Model-level result: Model Response records Exact 516 / >=516 gpt-5.5 75,401 44.0% gpt-5.4 25,214 19.8% gpt-5.2 247,575 0.34% gpt-5.3-codex 13,333 0.0% gpt-5.3-codex-spark 26,179 0.0% Monthly exact-516 clustering increased sharply: Month Exact 516 / >=516 Feb 2026 0.11% Mar 2026 2.45% Apr 2026 4.25% May 2026 53.30% Jun 2026 35.84% At the same time, overall reasoning-token intensity decreased: Month Mean reasoning tokens P90 reasoning tokens Feb 2026 268.1 772 Mar 2026 256.8 723 Apr 2026 228.7 669 May 2026 106.9 344 Jun 2026 168.5 515 Why this looks suspicious The anomaly is not simply higher reasoning-token usage overall. Mean and P90 reasoning-token intensity fell from February-April to May-June, while exact-516 clustering rose sharply. The clustering is also not evenly distributed across models. gpt-5.5 accounts for only 19.3% of responses but 82.0% of exact-516 events. Its exact-516 / >=516 ratio is about 33.6x higher than the non-GPT-5.5 baseline. The fixed values are also notable: 516, 1034, and 1552 look like repeated threshold boundaries rather than a naturally varying reasoning-token distribution. Expected behavior Reasoning-token counts for complex Codex tasks should vary naturally with task complexity and should not disproportionately cluster at exact fixed values for one model family. Actual behavior gpt-5.5 responses cluster heavily at exactly 516 reasoning tokens, with related spikes around 1034 and 1552. This pattern is much weaker or absent in several other models. Ask Could the Codex team investigate whether gpt-5.5 has a reasoning-budget, routing, truncation, fallback, or scheduler behavior that causes responses to terminate around 516/1034/1552 reasoning tokens? If this is expected behavior, it would be useful to know whether exact 516 indicates a normal stopping point, a budget cap, a degraded tier, or another internal threshold. Useful internal validation checks: Query token_count events with reasoning_output_tokens by model. Compare exact-value counts for 0, 516, 1034, and 1552. Compute count(reasoning_output_tokens = 516) / count(reasoning_output_tokens >= 516) by model and day. Compare gpt-5.5 against gpt-5.2, gpt-5.4, and Codex-specific variants. Replay matched complex tasks across GPT-5.2 and GPT-5.5 with quality evals, especially separating exact-516 responses from longer-reasoning responses. Reactions are currently unavailable Metadata Metadata Assignees No one assigned Labels bug Something isn't working Something isn't working model-behavior Issues related to behaviors exhibited by the model Issues related to behaviors exhibited by the model rate-limits Issues related to rate limits, quotas, and token usage reporting Issues related to rate limits, quotas, and token usage reporting Type No type Fields No fields configured for issues without a type. Projects No projects Milestone No milestone Relationships None yet Development No branches or pull requests Issue actions Open in GitHub Copilot app --- Codexチームは、gpt-5.5に推論バジェット、ルーティング、トランケーション、フォールバック、またはスケジューラのような挙動があり、レスポンスが約516/1034/1552の推論トークンで終了する原因となっているかどうかを調査してもらえないでしょうか？ もしこれが予期された挙動である場合、正確な516が通常の停止点、バジェットキャップ、パフォーマンス低下したティア、またはその他の内部しきい値を示しているのかを知ることが有益でしょう。 有用な内部検証チェック: モデルごとのreasoning_output_tokensを持つtoken_countイベントをクエリする。 0、516、1034、1552の正確な値のカウントを比較する。 モデルごとおよび日ごとにcount(reasoning_output_tokens = 516) / count(reasoning_output_tokens >= 516)を計算する。 gpt-5.5をgpt-5.2、gpt-5.4、およびCodex固有のバリアントと比較する。 品質評価を用いて、特に正確な516レスポンスとより長い推論レスポンスを分離して、gpt-5.2とgpt-5.5間で一致した複雑なタスクをリプレイする。