Side-by-side comparison · models
GLM 5.2 vs GPT OSS 20B
What we have checked about each, line by line. Every value shows where it came from and when we last confirmed it.
Models
GLM 5.2
Flagship GLM model for hybrid reasoning, coding, and agentic engineering.
Models
GPT OSS 20B
Open-weight GPT model for self-hosted reasoning and instruction-following workloads.
Checked facts
Line by line
| Measure | GLM 5.2 | GPT OSS 20B |
|---|---|---|
| Input price | $0.0070 per million tokens | $0.030 per million tokens |
| Output price | $6.80 per million tokens | $0.130 per million tokens |
| Price on ambient | $0.150 per million cached input tokens | Not checked yet |
| Price on ambient | $0.600 per million input tokens | Not checked yet |
| Price on ambient | $2.00 per million output tokens | Not checked yet |
| Price on atlas-cloud | $0.174 per million cached input tokens | Not checked yet |
| Price on atlas-cloud | $0.938 per million input tokens | Not checked yet |
| Price on atlas-cloud | $2.95 per million output tokens | Not checked yet |
| Price on baidu | $0.260 per million cached input tokens | Not checked yet |
| Price on baidu | $1.40 per million input tokens | Not checked yet |
| Price on baidu | $4.40 per million output tokens | Not checked yet |
| Price on baseten | $0.140 per million cached input tokens | Not checked yet |
| Price on baseten | $1.40 per million input tokens | Not checked yet |
| Price on baseten | $4.40 per million output tokens | Not checked yet |
| Price on cloudflare | $0.190 per million cached input tokens | Not checked yet |
| Price on cloudflare | $0.190 per million input tokens | Not checked yet |
| Price on cloudflare | $8.00 per million output tokens | Not checked yet |
| Price on coreweave | $0.140 per million cached input tokens | $0.030 per million cached input tokens |
| Price on coreweave | $0.620 per million input tokens | $0.030 per million input tokens |
| Price on coreweave | $2.42 per million output tokens | $0.130 per million output tokens |
| Price on decart | $0.480 per million cached input tokens | Not checked yet |
| Price on decart | $1.20 per million input tokens | Not checked yet |
| Price on decart | $8.00 per million output tokens | Not checked yet |
| Price on deepinfra | $0.084 per million cached input tokens | Not checked yet |
| Price on deepinfra | $0.450 per million input tokens | $0.030 per million input tokens |
| Price on deepinfra | $1.44 per million output tokens | $0.140 per million output tokens |
| Price on digitalocean | $0.105 per million cached input tokens | Not checked yet |
| Price on digitalocean | $0.700 per million input tokens | Not checked yet |
| Price on digitalocean | $2.20 per million output tokens | Not checked yet |
| Price on friendli | $0.260 per million cached input tokens | Not checked yet |
| Price on friendli | $1.40 per million input tokens | Not checked yet |
| Price on friendli | $4.40 per million output tokens | Not checked yet |
| Price on gmicloud | $0.260 per million cached input tokens | Not checked yet |
| Price on gmicloud | $1.40 per million input tokens | Not checked yet |
| Price on gmicloud | $4.40 per million output tokens | Not checked yet |
| Price on inceptron | $0.260 per million cached input tokens | Not checked yet |
| Price on inceptron | $1.40 per million input tokens | Not checked yet |
| Price on inceptron | $4.40 per million output tokens | Not checked yet |
| Price on mistralai | $0.140 per million cached input tokens | Not checked yet |
| Price on mistralai | $1.40 per million input tokens | Not checked yet |
| Price on mistralai | $4.40 per million output tokens | Not checked yet |
| Price on morph | $0.356 per million cached input tokens | Not checked yet |
| Price on morph | $1.05 per million input tokens | Not checked yet |
| Price on morph | $2.15 per million output tokens | Not checked yet |
| Price on nebius | $0.140 per million cached input tokens | Not checked yet |
| Price on nebius | $1.40 per million input tokens | Not checked yet |
| Price on nebius | $4.40 per million output tokens | Not checked yet |
| Price on novita | $0.121 per million cached input tokens | Not checked yet |
| Price on novita | $0.650 per million input tokens | Not checked yet |
| Price on novita | $2.04 per million output tokens | Not checked yet |
| Price on parasail | $0.260 per million cached input tokens | $0.020 per million cached input tokens |
| Price on parasail | $1.40 per million input tokens | $0.030 per million input tokens |
| Price on akashml | Not checked yet | $0.0090 per million cached input tokens |
| Price on akashml | Not checked yet | $0.018 per million input tokens |
| Price on akashml | Not checked yet | $0.090 per million output tokens |
| Price on amazon-bedrock | Not checked yet | $0.070 per million input tokens |
| Price on amazon-bedrock | Not checked yet | $0.150 per million output tokens |
| Price on darkbloom | Not checked yet | $0.0090 per million cached input tokens |
| Price on darkbloom | Not checked yet | $0.018 per million input tokens |
| Price on darkbloom | Not checked yet | $0.090 per million output tokens |
| Price on dekallm | Not checked yet | $0.029 per million cached input tokens |
| Price on dekallm | Not checked yet | $0.029 per million input tokens |
| Price on dekallm | Not checked yet | $0.140 per million output tokens |
| Price on google-vertex | Not checked yet | $0.070 per million input tokens |
| Price on google-vertex | Not checked yet | $0.250 per million output tokens |
| Price on groq | Not checked yet | $0.037 per million cached input tokens |
| Price on groq | Not checked yet | $0.075 per million input tokens |
| Price on groq | Not checked yet | $0.300 per million output tokens |
| Price on parasail | Not checked yet | $0.150 per million output tokens |
| Price on prime-intellect | Not checked yet | $0.021 per million cached input tokens |
| Price on prime-intellect | Not checked yet | $0.023 per million input tokens |
| Price on prime-intellect | Not checked yet | $0.034 per million output tokens |
| Price on siliconflow | Not checked yet | $0.040 per million input tokens |
| Price on siliconflow | Not checked yet | $0.180 per million output tokens |
| Latest version | z-ai/glm-5.2-20260616 | openai/gpt-oss-20b |
| Time to first token (ms) | 1,120 | 820 |
| Time to first token (ms) on ambient | 1,563 | Not checked yet |
| Time to first token (ms) on atlas-cloud | 1,218 | Not checked yet |
| Time to first token (ms) on baidu | 2,141 | Not checked yet |
| Time to first token (ms) on baseten | 2,856 | Not checked yet |
| Time to first token (ms) on cloudflare | 1,499 | Not checked yet |
| Time to first token (ms) on coreweave | 612 | 185 |
| Time to first token (ms) on decart | 1,204 | Not checked yet |
| Time to first token (ms) on deepinfra | 3,470 | 316 |
| Time to first token (ms) on digitalocean | 518 | Not checked yet |
| Time to first token (ms) on friendli | 521 | Not checked yet |
| Time to first token (ms) on gmicloud | 3,125 | Not checked yet |
| Time to first token (ms) on inceptron | 1,241 | Not checked yet |
| Time to first token (ms) on mistralai | 1,335 | Not checked yet |
| Time to first token (ms) on morph | 668 | Not checked yet |
| Time to first token (ms) on nebius | 27,561 | Not checked yet |
| Time to first token (ms) on novita | 2,586 | Not checked yet |
| Time to first token (ms) on parasail | 1,139 | 1,656 |
| Time to first token (ms) on phala | 2,618 | Not checked yet |
| Time to first token (ms) on qwen | 1,028 | Not checked yet |
| Time to first token (ms) on relace | 1,405 | Not checked yet |
| Time to first token (ms) on siliconflow | 1,302 | 1,131 |
| Time to first token (ms) on streamlake | 1,534 | Not checked yet |
| Time to first token (ms) on together | 706 | Not checked yet |
| Time to first token (ms) on venice | 1,987 | Not checked yet |
| Time to first token (ms) on wafer | 702 | Not checked yet |
| Time to first token (ms) on z-ai | 1,958 | Not checked yet |
| Output speed (tokens a second) | 175.3 | 222.3 |
| Output speed (tokens a second) on ambient | 32.5 | Not checked yet |
| Output speed (tokens a second) on atlas-cloud | 60 | Not checked yet |
| Output speed (tokens a second) on baidu | 65 | Not checked yet |
| Output speed (tokens a second) on baseten | 31 | Not checked yet |
| Output speed (tokens a second) on cloudflare | 99 | Not checked yet |
| Output speed (tokens a second) on coreweave | 93 | 54 |
| Output speed (tokens a second) on decart | 277 | Not checked yet |
| Output speed (tokens a second) on deepinfra | 42 | 80 |
| Output speed (tokens a second) on digitalocean | 71.5 | Not checked yet |
| Output speed (tokens a second) on friendli | 49 | Not checked yet |
| Output speed (tokens a second) on gmicloud | 36 | Not checked yet |
| Output speed (tokens a second) on inceptron | 58.5 | Not checked yet |
| Time to first token (ms) on akashml | Not checked yet | 708 |
| Time to first token (ms) on amazon-bedrock | Not checked yet | 436 |
| Time to first token (ms) on darkbloom | Not checked yet | 633 |
| Time to first token (ms) on dekallm | Not checked yet | 2,267 |
| Time to first token (ms) on groq | Not checked yet | 376 |
| Time to first token (ms) on prime-intellect | Not checked yet | 727 |
| Output speed (tokens a second) on akashml | Not checked yet | 96 |
| Output speed (tokens a second) on amazon-bedrock | Not checked yet | 310 |
| Output speed (tokens a second) on darkbloom | Not checked yet | 68 |
| Output speed (tokens a second) on dekallm | Not checked yet | 71 |
| Output speed (tokens a second) on groq | Not checked yet | 326 |
| Output speed (tokens a second) on parasail | Not checked yet | 71 |
| Output speed (tokens a second) on prime-intellect | Not checked yet | 96 |
| Output speed (tokens a second) on siliconflow | Not checked yet | 31 |
| Tokens served on OpenRouter in a day | 166.2B (for 10 Nov 2026) | 25.2B (for 10 Nov 2026) |
| Tool calls on OpenRouter in a day | 2.2M (for 10 Nov 2026) | 664.4K (for 10 Nov 2026) |
| Artificial Analysis Intelligence Index | 33.7 | 10 |
| Artificial Analysis Coding Index | 68.8 | 20.7 |
| Artificial Analysis Agentic Index | 38.4 | 1.2 |
| LMArena text rating (Elo) | 1,469 | 1,288 |
| LMArena web development rating (Elo) | 1,608 | Not checked yet |
| Design Arena rating (Elo) | 1,289 | 933 |
| Gpqa diamond | 91.9% (for 16 June 2026) | 60.8% (for 5 Aug 2025) |
| Swe bench verified | 78.7% (for 16 June 2026) | Not checked yet |
| Frontiermath tiers 1 3 v2 private | 59.2% (for 16 June 2026) | Not checked yet |
| Simpleqa verified | 34.2% (for 16 June 2026) | Not checked yet |
| Arc agi 2 | 22.8% (for 16 June 2026) | Not checked yet |
| Apex agents | 45.2% (for 16 June 2026) | Not checked yet |
| Arc agi | 77% (for 16 June 2026) | Not checked yet |
| Chess puzzles | 22% (for 16 June 2026) | Not checked yet |
| Dtbench | 93.6% (for 16 June 2026) | 68% (for 5 Aug 2025) |
| Ebr bench | 9.5% (for 16 June 2026) | Not checked yet |
| Frontiercode | 24.5% (for 16 June 2026) | Not checked yet |
| Frontiermath tier 4 v2 private | 29.3% (for 16 June 2026) | Not checked yet |
| Lmca | 45.8% (for 16 June 2026) | 14.5% (for 5 Aug 2025) |
| Mystery game puzzles | 19% (for 16 June 2026) | Not checked yet |
| Posttrainbench | 31.7% (for 16 June 2026) | Not checked yet |
| Proofbench | 35% (for 16 June 2026) | Not checked yet |
| Simplebench | 58.8% (for 16 June 2026) | Not checked yet |
| Surface evolver bench | 55.6% (for 16 June 2026) | Not checked yet |
| Weirdml | 70.1% (for 16 June 2026) | 40.9% (for 5 Aug 2025) |
| Otis mock aime 2024 2025 | 86.4% (saturated benchmark: the top models all score near the ceiling) (for 16 June 2026) | 65.3% (saturated benchmark: the top models all score near the ceiling) (for 5 Aug 2025) |
| Terminal bench | Not checked yet | 3.4% (for 5 Aug 2025) |
A line appears only when at least one record has a value we confirmed from a source. Values we have not confirmed are left out, never guessed. Records of different types share fewer lines.
Add a third
Compare with one more record
How this works
Share it, or build your own
This page works for any two or three records in the directory, of any type. The link stays the same whichever order you name them in, so you can share it as it is. Change the records in the address, for example /directories/vs/claude-code-vs-cursor, or start from our editorial comparisons.