Codex

Codex Radar

中文
📣📣📣 Notice

Important: Codex's official two-week referral reset-card redemption campaign ends at 00:00 on June 25 in UTC-12. Speedrunners who have not redeemed yet should do it soon.Calculating countdown

Links below

Speedrun Group Score History
Codex radar overview: four IQ Drop Radar cards, community experience, reset status, and official reset evidence

Quota Radar 6月24日12:58更新

Tier 5h 7d Basis
20x Pro $284.57 $1,707.42 measured
5x Pro $71.14 $426.86 estimated
Plus $14.23 $85.37 estimated
Trend $283.93 → $284.57 (+0.64)
$238 $277 $317 2026-06-19 20x Pro 5h $281.89 2026-06-20 20x Pro 5h $249.85 2026-06-21 20x Pro 5h $304.80 2026-06-22 20x Pro 5h $303.36 2026-06-23 20x Pro 5h $276.66 2026-06-24-am 20x Pro 5h $283.93 2026-06-24-pm 20x Pro 5h $284.57 6.19 6.20 6.21 6.22 6.23 6.24_am 6.24_pm

Using the latest valid 5h calibration (2026-06-24-pm). Measured with a 20x Pro US account. Background automations and other foreground tasks were paused before running the fixed task set; the estimate converts the 5h quota change before/after completion and the tokens consumed. Formula: 20x Pro 5h = $119.52 / (43% - 1%) * 100 = $284.57; same-token basis ≈ 144.97M / (43% - 1%) * 100 = 345.18M.

IQ Drop Radar 6月24日12:58更新

GPT-5.5-xhigh 125.0
GPT-5.5-high 87.5
GPT-5.5-medium 75.0
GPT-5.4-xhigh 87.5

GPT-5.5-xhigh: 6.24_pm, 10/12 passed; baseline 8/12

IQ index 120 100 80 60 90-110 band 6.17_pm GPT-5.5 xhigh: IQ index 87.5, 7/12, Cost $46.38, Time 170 min, cache hit 94.2% 6.18 GPT-5.5 xhigh: IQ index 125.0, 10/12, Cost $40.58, Time 165 min, cache hit 95.0% 6.19 GPT-5.5 xhigh: IQ index 100.0, 8/12, Cost $46.74, Time 193 min, cache hit 95.7% 6.20 GPT-5.5 xhigh: IQ index 75.0, 6/12, Cost $42.61, Time 160 min, cache hit 95.0% 6.21 GPT-5.5 xhigh: IQ index 87.5, 7/12, Cost $38.98, Time 153 min, cache hit 94.0% 6.22_am GPT-5.5 xhigh: IQ index 100.0, 8/12, Cost $42.99, Time 171 min, cache hit 93.3% 6.22_pm GPT-5.5 xhigh: IQ index 50.0, 4/12, Cost $44.09, Time 199 min, cache hit 94.6% 6.23 GPT-5.5 xhigh: IQ index 125.0, 10/12, Cost $40.21, Time 158 min, cache hit 95.3% 6.24_am GPT-5.5 xhigh: IQ index 87.5, 7/12, Cost $37.26, Time 144 min, cache hit 93.6% 6.24_pm GPT-5.5 xhigh: IQ index 125.0, 10/12, Cost $40.38, Time 154 min, cache hit 94.2% 6.17_pm GPT-5.5 high: IQ index 87.5, 7/12, Cost $32.71, Time 161 min, cache hit 92.2% 6.18 GPT-5.5 high: IQ index 100.0, 8/12, Cost $29.86, Time 113 min, cache hit 93.0% 6.19 GPT-5.5 high: IQ index 87.5, 7/12, Cost $30.15, Time 119 min, cache hit 94.1% 6.20 GPT-5.5 high: IQ index 87.5, 7/12, Cost $28.20, Time 104 min, cache hit 92.2% 6.21 GPT-5.5 high: IQ index 87.5, 7/12, Cost $26.58, Time 103 min, cache hit 94.1% 6.22_am GPT-5.5 high: IQ index 62.5, 5/12, Cost $26.59, Time 110 min, cache hit 93.8% 6.22_pm GPT-5.5 high: IQ index 62.5, 5/12, Cost $30.53, Time 138 min, cache hit 93.7% 6.23 GPT-5.5 high: IQ index 87.5, 7/12, Cost $31.42, Time 116 min, cache hit 93.5% 6.24_am GPT-5.5 high: IQ index 100.0, 8/12, Cost $29.01, Time 109 min, cache hit 93.2% 6.24_pm GPT-5.5 high: IQ index 87.5, 7/12, Cost $29.20, Time 106 min, cache hit 94.7% 6.17_pm GPT-5.5 medium: IQ index 75.0, 6/12, Cost $24.14, Time 117 min, cache hit 90.7% 6.18 GPT-5.5 medium: IQ index 75.0, 6/12, Cost $22.53, Time 88 min, cache hit 91.7% 6.19 GPT-5.5 medium: IQ index 87.5, 7/12, Cost $21.32, Time 96 min, cache hit 94.1% 6.20 GPT-5.5 medium: IQ index 75.0, 6/12, Cost $22.50, Time 82 min, cache hit 93.8% 6.21 GPT-5.5 medium: IQ index 75.0, 6/12, Cost $22.64, Time 84 min, cache hit 93.0% 6.22_am GPT-5.5 medium: IQ index 75.0, 6/12, Cost $23.49, Time 92 min, cache hit 93.0% 6.22_pm GPT-5.5 medium: IQ index 87.5, 7/12, Cost $26.30, Time 106 min, cache hit 91.5% 6.23 GPT-5.5 medium: IQ index 87.5, 7/12, Cost $22.27, Time 155 min, cache hit 92.8% 6.24_am GPT-5.5 medium: IQ index 87.5, 7/12, Cost $22.21, Time 83 min, cache hit 93.4% 6.24_pm GPT-5.5 medium: IQ index 75.0, 6/12, Cost $25.43, Time 90 min, cache hit 93.2% 6.17_pm GPT-5.4 xhigh: IQ index 62.5, 5/12, Cost $23.84, Time 190 min, cache hit 95.9% 6.18 GPT-5.4 xhigh: IQ index 100.0, 8/12, Cost $24.23, Time 202 min, cache hit 96.2% 6.19 GPT-5.4 xhigh: IQ index 75.0, 6/12, Cost $23.01, Time 208 min, cache hit 96.3% 6.20 GPT-5.4 xhigh: IQ index 62.5, 5/12, Cost $26.61, Time 218 min, cache hit 97.2% 6.21 GPT-5.4 xhigh: IQ index 62.5, 5/12, Cost $21.53, Time 199 min, cache hit 96.0% 6.22_am GPT-5.4 xhigh: IQ index 62.5, 5/12, Cost $22.21, Time 234 min, cache hit 95.5% 6.22_pm GPT-5.4 xhigh: IQ index 62.5, 5/12, Cost $22.89, Time 241 min, cache hit 96.0% 6.23 GPT-5.4 xhigh: IQ index 87.5, 7/12, Cost $22.39, Time 239 min, cache hit 95.0% 6.24_am GPT-5.4 xhigh: IQ index 62.5, 5/12, Cost $25.10, Time 190 min, cache hit 96.4% 6.24_pm GPT-5.4 xhigh: IQ index 87.5, 7/12, Cost $24.51, Time 200 min, cache hit 96.0% 6.17_pm 6.18 6.19 6.20 6.21 6.22_am 6.22_pm 6.23 6.24_am 6.24_pm 5.5-xhigh 5.5-high 5.5-medium 5.4-xhigh
Metric5.5-xhigh5.5-high5.5-medium5.4-xhighGLM-5.2 max
Passes10/127/126/127/127/12
IQ125.087.575.087.587.5
Agent steps470421407545959
Cost$40.38$29.20$25.43$24.51$31.99
Cache hit rate94.2%94.7%93.2%96.0%97.5%
Time154 min106 min90 min200 min445 min
Total tokens39.09M30.60M25.08M50.20M88.81M

Note: GLM-5.2 is run irregularly for reference only; test setup: GLM Coding Plan Lite + CC.

Fixed DeepSWE task set

The daily IQ point is scored on this fixed mixed-language DeepSWE subset. Each row shows today's result for the tracked models.

12 tasks
  1. 01
    Add JSONPath query APIs to orderedmap and Starlark modules ytt-jsonpath-query-api
    GO history pass 77% (65/84)
    5.5-xhigh pass 5.5-high pass 5.5-medium pass 5.4-xhigh pass
  2. 02
    Add build-time grammar conflict analysis to participle participle-grammar-conflict-analysis
    GO history pass 19% (16/84)
    5.5-xhigh pass 5.5-high fail 5.5-medium fail 5.4-xhigh pass
  3. 03
    Harden module loading, cache introspection, and script flags abs-module-cache-flags
    GO history pass 90% (76/84)
    5.5-xhigh pass 5.5-high pass 5.5-medium pass 5.4-xhigh pass
  4. 04
    Add multipart response parsing to HTTPX httpx-multipart-response-parsing
    PYTHON history pass 94% (79/84)
    5.5-xhigh pass 5.5-high pass 5.5-medium pass 5.4-xhigh fail
  5. 05
    Add incremental cache controls to Bandit bandit-incremental-cache-control
    PYTHON history pass 36% (30/84)
    5.5-xhigh pass 5.5-high fail 5.5-medium fail 5.4-xhigh fail
  6. 06
    Add session bundle recording and replay to IPython ipython-session-bundle-replay
    PYTHON history pass 19% (16/84)
    5.5-xhigh pass 5.5-high fail 5.5-medium fail 5.4-xhigh fail
  7. 07
    Add a per-origin circuit breaker to ofetch ofetch-per-origin-circuit-breaker
    TYPESCRIPT history pass 95% (80/84)
    5.5-xhigh pass 5.5-high pass 5.5-medium pass 5.4-xhigh pass
  8. 08
    Add link format conversion between wiki and markdown syntax obsidian-linter-link-format-conversion
    TYPESCRIPT history pass 13% (11/84)
    5.5-xhigh fail 5.5-high fail 5.5-medium fail 5.4-xhigh fail
  9. 09
    Add atomic signal selectors to Kea kea-atomic-signal-selectors
    TYPESCRIPT history pass 65% (55/84)
    5.5-xhigh pass 5.5-high pass 5.5-medium fail 5.4-xhigh pass
  10. 10
    Add shorthand expansion and compression to the lexer csstree-shorthand-expansion-compression
    JAVASCRIPT history pass 51% (43/84)
    5.5-xhigh pass 5.5-high fail 5.5-medium pass 5.4-xhigh pass
  11. 11
    Add deterministic multi-key sorting to fd fd-deterministic-multi-key-sorting
    RUST history pass 31% (26/84)
    5.5-xhigh fail 5.5-high pass 5.5-medium fail 5.4-xhigh fail
  12. 12
    Preserve structure needed by stylesheet selectors oxvg-structural-selector-preservation
    RUST history pass 73% (61/84)
    5.5-xhigh pass 5.5-high pass 5.5-medium pass 5.4-xhigh pass

Community Score

Please rate only models you used in the last 24 hours to avoid adding noise. Thank you.

Reference: 9-10 clearly strong, 7-8 normally usable, 5-6 barely usable, 3-4 poor, 1-2 nearly unusable. Consider accuracy, rework, speed, stability, and quota/service impact.

Last 24h rating
GPT-5.5 xhigh
No rating in 24h
N/A 0 ratings
GPT-5.5 high
No rating in 24h
N/A 0 ratings
GPT-5.5 medium
No rating in 24h
N/A 0 ratings
GPT-5.4 xhigh
No rating in 24h
N/A 0 ratings

Loading ratings...

High-quality Codex technical community

WeChat public account QR code WeChat public account
Codex Radar WeChat group QR code WeChat Radar Group
Tibo Radar
Tibo's X avatar
--:-- -- Calculating
San Francisco Bay Area / PT
Public location estimate: 55% The supplied public posts do not self-disclose a temporary location or travel, so the default PT timezone is used. Source ↗
API

JSON is current state; RSS focuses on official windows and reset alerts.

JSONCurrent state/current.json RSSOfficial window alerts/feed.xml