Side-by-side comparison · models
Qwen3.8 2.4T A95B vs Qwen3.8 27B (free)
What we have checked about each, line by line. Every value shows where it came from and when we last confirmed it.
Models
Qwen3.8 2.4T A95B
Qwen3.8 2.4T A95B is an open-weight sparse mixture-of-experts model from Qwen and the open-weight variant of Qwen3.8 Max, with 95 billion active parameters out.
Models
Qwen3.8 27B (free)
Qwen3.8 27B is an open-weight dense vision-language model from Qwen.
Checked facts
Line by line
| Measure | Qwen3.8 2.4T A95B | Qwen3.8 27B (free) |
|---|---|---|
| Input price | $2.00 per million tokens | Not checked yet |
| Output price | $6.00 per million tokens | Not checked yet |
| Price on novita | $0.250 per million cached input tokens | $0.085 per million cached input tokens |
| Price on novita | $2.00 per million input tokens | $0.420 per million input tokens |
| Price on novita | $6.00 per million output tokens | $3.00 per million output tokens |
| Price on qwen | $0.250 per million cached input tokens | Not checked yet |
| Price on qwen | $2.00 per million input tokens | Not checked yet |
| Price on qwen | $6.00 per million output tokens | Not checked yet |
| Price on siliconflow | $0.250 per million cached input tokens | Not checked yet |
| Price on siliconflow | $2.00 per million input tokens | Not checked yet |
| Price on siliconflow | $6.00 per million output tokens | Not checked yet |
| Price on together | $0.250 per million cached input tokens | Not checked yet |
| Price on together | $2.00 per million input tokens | Not checked yet |
| Price on together | $6.00 per million output tokens | Not checked yet |
| Price on venice | $0.250 per million cached input tokens | Not checked yet |
| Price on venice | $2.00 per million input tokens | Not checked yet |
| Price on venice | $6.00 per million output tokens | Not checked yet |
| Price on akashml | Not checked yet | $0.050 per million cached input tokens |
| Price on akashml | Not checked yet | $0.225 per million input tokens |
| Price on akashml | Not checked yet | $1.98 per million output tokens |
| Price on cerebras | Not checked yet | $0.990 per million cached input tokens |
| Price on cerebras | Not checked yet | $0.990 per million input tokens |
| Price on cerebras | Not checked yet | $1.49 per million output tokens |
| Price on chutes | Not checked yet | $0.0039 per million cached input tokens |
| Price on chutes | Not checked yet | $0.039 per million input tokens |
| Price on chutes | Not checked yet | $1.25 per million output tokens |
| Price on cloudflare | Not checked yet | $0.050 per million cached input tokens |
| Price on cloudflare | Not checked yet | $0.450 per million input tokens |
| Price on cloudflare | Not checked yet | $3.20 per million output tokens |
| Price on coreweave | Not checked yet | $0.150 per million cached input tokens |
| Price on coreweave | Not checked yet | $0.400 per million input tokens |
| Price on coreweave | Not checked yet | $3.00 per million output tokens |
| Price on darkbloom | Not checked yet | $0.018 per million cached input tokens |
| Price on darkbloom | Not checked yet | $0.035 per million input tokens |
| Price on darkbloom | Not checked yet | $1.50 per million output tokens |
| Price on deepinfra | Not checked yet | $0.037 per million cached input tokens |
| Price on deepinfra | Not checked yet | $0.150 per million input tokens |
| Price on deepinfra | Not checked yet | $1.88 per million output tokens |
| Price on dekallm | Not checked yet | $0.020 per million cached input tokens |
| Price on dekallm | Not checked yet | $0.040 per million input tokens |
| Price on dekallm | Not checked yet | $2.50 per million output tokens |
| Price on gmicloud | Not checked yet | $0.250 per million cached input tokens |
| Price on gmicloud | Not checked yet | $0.450 per million input tokens |
| Price on gmicloud | Not checked yet | $3.20 per million output tokens |
| Price on ionstream | Not checked yet | $0.020 per million cached input tokens |
| Price on ionstream | Not checked yet | $0.040 per million input tokens |
| Price on ionstream | Not checked yet | $1.95 per million output tokens |
| Price on mancer | Not checked yet | $0.075 per million cached input tokens |
| Price on mancer | Not checked yet | $0.150 per million input tokens |
| Price on mancer | Not checked yet | $1.25 per million output tokens |
| Price on modal | Not checked yet | $0.020 per million cached input tokens |
| Price on modal | Not checked yet | $0.040 per million input tokens |
| Price on modal | Not checked yet | $2.30 per million output tokens |
| Price on modelrun | Not checked yet | $0.090 per million cached input tokens |
| Price on modelrun | Not checked yet | $0.200 per million input tokens |
| Price on modelrun | Not checked yet | $2.50 per million output tokens |
| Price on near-ai | Not checked yet | $0.018 per million cached input tokens |
| Price on near-ai | Not checked yet | $0.040 per million input tokens |
| Price on near-ai | Not checked yet | $1.35 per million output tokens |
| Price on nebius | Not checked yet | $0.100 per million cached input tokens |
| Price on nebius | Not checked yet | $0.450 per million input tokens |
| Price on nebius | Not checked yet | $3.00 per million output tokens |
| Price on parasail | Not checked yet | $0.050 per million cached input tokens |
| Price on parasail | Not checked yet | $0.240 per million input tokens |
| Latest version | qwen/qwen3.8-2.4t-a95b-20260812 | qwen/qwen3.8-27b-20260814 |
| Time to first token (ms) | 2,870 | 3,920 |
| Time to first token (ms) on novita | 3,882 | 2,908 |
| Time to first token (ms) on qwen | 1,712 | 3,865 |
| Time to first token (ms) on siliconflow | 2,808 | Not checked yet |
| Time to first token (ms) on together | 5,767 | Not checked yet |
| Time to first token (ms) on venice | 1,202 | 533 |
| Output speed (tokens a second) | 39.8 | 41.3 |
| Output speed (tokens a second) on novita | 33 | Not checked yet |
| Output speed (tokens a second) on qwen | 38 | Not checked yet |
| Output speed (tokens a second) on siliconflow | 34 | Not checked yet |
| Output speed (tokens a second) on together | 21 | Not checked yet |
| Output speed (tokens a second) on venice | 139 | Not checked yet |
| Time to first token (ms) on akashml | Not checked yet | 683 |
| Time to first token (ms) on cerebras | Not checked yet | 449 |
| Time to first token (ms) on chutes | Not checked yet | 2,307 |
| Time to first token (ms) on cloudflare | Not checked yet | 378 |
| Time to first token (ms) on coreweave | Not checked yet | 976 |
| Time to first token (ms) on darkbloom | Not checked yet | 3,222 |
| Time to first token (ms) on deepinfra | Not checked yet | 948 |
| Time to first token (ms) on dekallm | Not checked yet | 639 |
| Time to first token (ms) on gmicloud | Not checked yet | 3,085 |
| Time to first token (ms) on ionstream | Not checked yet | 1,255 |
| Time to first token (ms) on mancer | Not checked yet | 855 |
| Time to first token (ms) on modal | Not checked yet | 602 |
| Time to first token (ms) on modelrun | Not checked yet | 278 |
| Time to first token (ms) on near-ai | Not checked yet | 1,540 |
| Time to first token (ms) on nebius | Not checked yet | 797 |
| Time to first token (ms) on parasail | Not checked yet | 580 |
| Time to first token (ms) on phala | Not checked yet | 973 |
| Time to first token (ms) on prime-intellect | Not checked yet | 683 |
| Time to first token (ms) on reka | Not checked yet | 1,211 |
| Time to first token (ms) on streamlake | Not checked yet | 5,825 |
| Time to first token (ms) on tenstorrent | Not checked yet | 597 |
| Time to first token (ms) on wafer | Not checked yet | 834 |
| Output speed (tokens a second) on akashml | Not checked yet | 41 |
| Output speed (tokens a second) on cerebras | Not checked yet | 624.5 |
| Output speed (tokens a second) on chutes | Not checked yet | 34 |
| Output speed (tokens a second) on cloudflare | Not checked yet | 38.5 |
| Output speed (tokens a second) on coreweave | Not checked yet | 27 |
| Output speed (tokens a second) on darkbloom | Not checked yet | 22 |
| Output speed (tokens a second) on deepinfra | Not checked yet | 23 |
| Output speed (tokens a second) on dekallm | Not checked yet | 119 |
| Output speed (tokens a second) on gmicloud | Not checked yet | 45 |
| Output speed (tokens a second) on ionstream | Not checked yet | 52 |
| Output speed (tokens a second) on mancer | Not checked yet | 65 |
| Output speed (tokens a second) on modal | Not checked yet | 111 |
| Output speed (tokens a second) on modelrun | Not checked yet | 165 |
| Tokens served on OpenRouter in a day | 3.4B (for 10 Nov 2026) | Not checked yet |
| Tool calls on OpenRouter in a day | 40.8K (for 10 Nov 2026) | Not checked yet |
| Artificial Analysis Intelligence Index | 39.9 | Not checked yet |
| Artificial Analysis Coding Index | 71.9 | Not checked yet |
| Artificial Analysis Agentic Index | 50.1 | Not checked yet |
| Arc agi 2 | Not checked yet | 42.4% (for 14 Aug 2026) |
| Apex agents | Not checked yet | 47.5% (for 14 Aug 2026) |
| Arc agi | Not checked yet | 87.5% (for 14 Aug 2026) |
| Dtbench | Not checked yet | 88% (for 14 Aug 2026) |
| Lmca | Not checked yet | 41.4% (for 14 Aug 2026) |
| Proofbench | Not checked yet | 16% (for 14 Aug 2026) |
| Surface evolver bench | Not checked yet | 45% (for 14 Aug 2026) |
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.